DP
RIETI Discussion Paper Series 05-E-004
Why Did Japan's TFP Growth Slow Down in the Lost Decade?
An Empirical Analysis Based on Firm-Level Data of Manufacturing Firms
FUKAO Kyoji
RIETI
KWON Hyeog Ug
Hitotsubashi University
The Research Institute of Economy, Trade and Industry
http://www.rieti.go.jp/en/
RIETI Discussion Paper Series 05-E-004
Why Did Japan’s TFP Growth Slow Down in the Lost Decade?
An Empirical Analysis Based on Firm-Level Data of Manufacturing Firms*
Kyoji Fukao**
Hitotsubashi Universitya and RIETIb
Hyeog Ug Kwon
Hitotsubashi Universityc
Revised Version
February 2005
*We are grateful to the discussant, Kiyohiko G. Nishimura, and participants of the CIRJE-TCER
Macro Conference for their comments and suggestions on a preliminary version of this paper.
**Correspondence: Kyoji Fukao, Institute of Economic Research, Hitotsubashi University, Naka 2-1,
Kunitachi, Tokyo 186 JAPAN. Tel.: +81-42-580-8359, Fax: +81-42-580 -8333, e-mail:
k.fukao@srv.cc.hit-u.ac.jp
a
Professor.
b
Faculty fellow.
c
Research Fellow.
1. Introduction
For more than a decade Japan has experienced a phase of unprecedentedly slow growth. The
causes of this stagnation are the subject of considerable controversy. One group of scholars attributes
the disappointing performance to a lack of effective demand and a liquidity trap caused by deflation.1
The other group points out that there are several important supply-side factors, which reduced Japan’s
economic growth. For example, Japan’s aging population and a gradual reduction in the statutory
work-week have contributed to a slowdown in the growth of labor input.2 Japan also experienced a
decline in total factor productivity (TFP), which has important effects on economic growth not only
because it reduces output growth by itself but also because it diminishes the rate of return to capital
and discourages private investment.
Although there are as many different estimates for Japan’s recent TFP growth as there are
studies on this issue, most economists seem to agree that Japan’s TFP growth substantially declined in
the 1990s. Probably the most popular explanation of Japan’s TFP growth slowdown is the “zombie”
hypothesis. This states that in order to conceal their bad loans, Japanese banks have been keeping alive
money-losing large borrowers by “evergreening” loans and discounting lending rates, although the
chance that these borrowers will recover is slim (Caballero, Hoshi and Kashyap 2004). Because of the
existence of zombie firms, the entry and growth of more productive firms are impeded and TFP
growth slows down in industries infested by zombies (Ahearne and Shinada 2004). Japanese banks’
bad loans are concentrated in non-manufacturing sectors, such as real estate, construction, commerce,
and services, since a major cause of the bad loans is the burst of the land price bubble in the early
1990s. For example, according to Caballero, Hoshi, and Kashyap (2004) the share of total assets held
by zombie firms in the total assets held by publicly traded firms was around 10 percent in the
1
For example, see Yoshikawa (2003) and M. Fukao (2003).
2
On these supply side factors, see Hayashi and Prescott (2002).
1
manufacturing sector in the 1998-2002 period whereas it was around 30 percent in real estate and
services, and 20 percent in construction and retail and wholesale (excluding the nine largest general
trading companies). 3 Therefore, according to the zombie hypothesis, we would expect that the
slowdown of Japan’s TFP has been concentrated mainly in the non-manufacturing sector.
Contrary to this conjecture, the majority of recent studies on sectoral TFP growth have found
that the slowdown in TFP growth was more serious in the manufacturing than in the
non-manufacturing sector (Yoshikawa and Matsumoto 2001, Nishimura and Minetaki 2003,
Miyagawa 2003, and Fukao et al. 2004). Given that the slowdown in TFP growth has been more
severe in the manufacturing sector, there is a need for more detailed analysis of this trend and its
causes. The present paper aims to examine the issue using firm-level data of the Ministry of Economy,
International Trade and Industry’s Kigyo Katsudo Kihon Chosa (Basic Survey on Business Activities
by Enterprises), which cover most of Japan’s manufacturing activities for the period of 1994-2001.
In our analysis, we concentrate on two issues in particular. First, we decompose TFP growth in
the manufacturing sector into a within-firm effect, a reallocation effect, and an entry-exit effect. If
firms compete with each other and entry barriers are low, high-productivity firms will enter the market
and expand their production. This “metabolism” will enhance the TFP growth of the industry. Using
the same micro-data of the Kigyo Katsudo Kihon Chosa, for the 1994-1998 period, Nishimura, et al.
(2003) and Fukao and Kwon (2003) studied the productivity of firms and conducted productivity
decompositions. In spite of the difference of the methodology adopted,4 both studies found that the
3
We should note that their dataset, which is taken from the Development Bank of Japan Database
covers a very limited percentage of economic activity in Japan. In the case of the non-manufacturing
sectors, the coverage measured by the percentage of workers is less than 10%. We will discuss this
issue in more detail in section 4.
4
Nishihmura et al. adopted the methodology used by Griliches and Regev (1995) and Aw, Chen,
and Roberts (2001). However, this methodology makes it difficult to separate the entry and exit
effects and to identify the covariance effect (the definition of these effects is presented in section 3).
2
average TFP level of exiting firms was higher than that of staying firms in some industries. This
“negative metabolism” may have slowed down TFP growth in the manufacturing sector. In this paper,
using the updated data, we decompose TFP growth for the longer period of 1994-2001. By adopting
the methodology used by Baily et al. (1992) and Forster et al. (1998), which has been commonly used
in recent studies, we can compare our results with preceding studies on the US, a number of European
countries, and Korea.
Second, we measure the gap in the TFP level between a group of high–TFP firms and a group of
low–TFP firms and compare the characteristics of these two groups. We show that the TFP gap
between the two groups is widening in many industries, including drugs and medicine, electronic data
processing machines and electronic equipment, and motor vehicles, where R&D intensity is high and
the internationalization of firms is more advanced (we measure internationalization by outward direct
investment, the introduction of foreign capital, and procurements from abroad). We found that the
high–TFP firms tend to have a higher R&D intensity, a higher degree of internationalization, larger
scale, and a lower liability-asset ratio.
We also found that greater R&D intensity and internationalization have positive effects on
firms’ TFP growth. Similar to the “IT (information technology) divide” among workers, a new divide
caused by R&D and internationalization seems to be emerging and growing in Japan’s manufacturing
industry.5 We also show that the reduction of the number of workers by the low–TFP firms is not much
Fukao and Kwon adopted the methodology used by Baily, Hulten and Campbell (1992) and Forster,
Haltiwanger and Krizan (1998), which allows us to separate the entry and exit effects and identify
the covariance effect. Another difference is that Nishimura et al. used value added as a measure of
output whereas Fukao and Kwon used gross output as a measure of output.
5
In several years, the METI Survey included questions on firms’ introduction of information
technology, such as the introduction of CAD (computer-aided design) and CAM (computer-aided
manufacturing) systems and the use of LANs (local area networks) within a firm group. But the
questions changed over time and the response ratio was not high. If we used these data in our
3
larger than the reduction by the high–TFP firms. The sales growth of the high–TFP firms is smaller
than that of the low-TFP firms. Most high-TFP firms are also reducing their employment, probably
because of organizational restructuring and the relocation of production abroad. This finding suggests
that the “metabolism” is not working well in Japan’s manufacturing sector.
The paper is organized as follows: in the next section, we present an overview of studies on
Japan’s TFP growth in the 1990s at the macro- and the industry-level. In section 3 we conduct a
decomposition of TFP growth in the manufacturing sector and compare our results with preceding
studies on other developed economies. In section 4, we examine the movement in the TFP level gap
between the 75 percentile firm and the 25 percentile firm by industry and by year and show that the
gap has widened in major industries. We also compare the characteristics, TFP growth, and growth
rate of employment of the top firm-group and the bottom firm-group. In section 5, we summarize our
results and discuss the policy implications of our findings.
2. Did TFP Growth Really Slow Down in the 1990s?
In this section we present a brief overview of the empirical research on Japan’s TFP growth rate
at the macro- and sectoral-levels in the 1990s. Table 2.1 summarizes the major results of preceding
studies and compares the methodologies adopted. We placed studies with more pessimistic results (a
large decline in TFP growth during the 1990s when compared with the 1980s) at the top of the table
and placed those with more optimistic results at the bottom of the table.
Insert Table 2.1
regression analysis, our sample size would be drastically reduced. For this reason, we do not analyze
the effect of information technology on TFP growth in this paper.
4
The table shows that there are substantial differences among the studies in the estimated decline
in TFP growth from the 1980s to the 1990s. Hayashi and Prescott (2002) and Yoshikawa and
Matsumoto (2001) obtained the most pessimistic results, suggesting that the TFP growth rate at the
macro-level declined by more than 2 percentage points from the 1980s to 1990s. According to the
neoclassical growth model, the decline in TFP growth will also reduce the equilibrium growth rate of
the real capital stock in balanced growth. If we assume a Cobb-Douglas production function with a
capital share of one-third, a 2 percentage-point decline in TFP growth will cause a 3 percentage-point
(=2+2/3) decline in the balanced growth rate. In contrast with these pessimistic results, several studies,
such as Jorgenson and Motohashi (2003) and Kawamoto (2004), found that the TFP growth in the
1990s was not substantially smaller than in the 1980s.
If we carefully compare the methodologies adopted and the datasets used, we can see what
causes the great variation in results and try to obtain a more accurate measure of TFP trends.6 For
example, Hayashi and Prescott (2002) do not take account of changes in the quality of labor. As the
improvement in the quality of labor in Japan has slowed in recent years, their study overestimates the
decline in TFP growth by neglecting such changes. Moreover, they do not take account of changes in
capacity utilization. This factor also contributed to their overestimation of the decline in TFP growth.7
6
Such a comparison is provided by Inui and Kwon (2004), who used the original datasets for factor
inputs, gross outputs, and income and cost shares used in the studies by Hayashi and Prescott (2002),
Jorgenson and Motohashi (2003), Cabinet Office (2002), and Fukao et al. (2003) and examined what
differences in the datasets and methodologies were responsible for the large discrepancies among the
results for TFP growth.
7
When the capital stock is not fully utilized, the marginal productivity of capital might be different
from the cost of capital. As Morrison (1993) has shown, we can tackle this issue more rigorously by
estimating the variable cost function. Using micro data of Japanese manufacturing firms, Fukao and
Kwon (2004) estimated variable cost functions and made adjustments for capacity utilization and
scale economies. They found that the rate of technological progress, which is defined as a downward
shift of the variable cost function, declined from 1994 to 2001 in many manufacturing sectors.
5
Finally, in their growth accounting they use real GNP instead of GDP as an output measure and include
Japan’s net external assets in the capital stock. In GNP statistics, the rate of return to domestic capital
is in gross terms and includes capital depreciation. On the other hand, the rate of return to Japan’s
external assets is recorded in net terms. Therefore, the appropriate capital cost of net external assets for
growth accounting is usually smaller than the cost of capital located in Japan. Hayashi and Prescott
(2002) did not take account of this difference and assumed that the cost share of capital was constant
over time. Since Japan accumulated a huge amount of net external assets in the 1990s,8 Hayashi and
Prescott seem to have overestimated the cost share of capital in the 1990s, the contribution of capital
deepening in the 1990s, and, as a result, the decline in Japan’s TFP growth from the 1980s to the
1990s.
A more optimistic result was obtained by Jorgenson and Motohashi (2003). They found that
Japan’s TFP growth rate declined from 0.96% in 1975-90 to 0.61% in 1990-95 but then accelerated
again to 1.04% in 1995-2000. This optimistic result is partly based on their assumption on the deflator
for information technology (IT) products. They assumed that the relative price of IT products
compared with non-IT products in Japan has declined in a similar fashion as in the US. They used their
own IT product deflator, calculated as ((US IT product price)/(US non-IT product price))*(Japan’s
non-IT product price), instead of Japan’s official statistics. Since the relative price of IT products
declined more drastically in the US than in Japan, this procedure raises their estimate of the GDP
growth rate and the TFP growth rate.9
Jorgenson and Motohashi adopt this procedure because they believe that quality improvements
8
From the end of 1991 to the end of 2001, Japan accumulated net external assets of 117.7 trillion
yen (Annual Report of National Accounts (various years), Economic and Social Research Institute,
Cabinet Office, Government of Japan.)
9
The lower price of IT products means larger IT investment. This factor reduces the estimate of the
TFP growth rate.
6
in IT products are not sufficiently taken into account in Japan’s price statistics.10 Although the authors
raised an important question, it seems brave to directly apply US relative prices to Japan. We need a
more rigorous analysis of the international price gap and the size of a hypothetical price decline, which
is equivalent to actual quality improvements in IT products.
There are many other differences in estimation procedures between Jorgenson and Motohashi’s
(2003) study and the other studies. Jorgenson and Motohashi explicitly treat land as a production
factor, but all the other studies neglect land input. The inclusion of land lowers the cost share of other
inputs. This difference makes their estimate of TFP growth higher than in the other studies. They also
include consumer durables and computer software in capital input, which most of the other studies do
not do.
Another important study with optimistic results is that of Kawamoto (2004). He found that the
TFP growth rate for Japan’s private sector in 1990-1998 was 1.9%, which is identical with the TFP
growth rate he obtained for 1980-1990. This optimistic result is mainly based on the following two
factors. First, following Basu and Kimball (1997), Kawamoto assumed that the sole cost of changing
the workweek of capital is a “shift premium” — firms need to pay higher wages to compensate
employees for working overtime — and used changes in hours per worker as a proxy for unobserved
changes in both labor effort and capital utilization. Second, he found that there are large diseconomies
of scale in the non-manufacturing sector. The estimated returns-to-scale coefficient in the
non-manufacturing sector was 0.65. Since the production share of the non-manufacturing sector
expanded rapidly in the 1990s, he attributed a substantial part of the productivity growth slowdown to
the diseconomies of scale instead of to the slowdown in technological progress.11
10
Colecchia and Schreyer (2002) adopted a similar approach in their comparative analysis of OECD
countries.
11
It is interesting to note that even Kawamoto (2004) finds that there was a considerable slowdown
in TFP growth in the durable manufacturing sector in the 1990s (see Table 2.1).
7
Despite the boldness of his assumption on the “shift premium” he does not validate its
applicability to the Japanese economy. Although his findings on the large diseconomies of scale in the
non-manufacturing sector seem to be inconsistent with the actual existence of large firms, he does not
confirm this finding using firm-level data. Kawamoto raised very interesting issues; but it seems that
we need more empirical research to verify his surprising results.
Table 2.1 also shows another important point, which we would like to stress. Many studies, such
as Yoshikawa and Matsumoto (2001), Nishimura and Minetaki (2003), Miyagawa (2003), Fukao et al.
(2004), and Kawamoto (2004) found that the slowdown in TFP growth in the manufacturing sector
was more severe than that in the non-manufacturing sector, even if we take account of changes in
capacity utilization.12 While recent studies on Japan’s economic stagnation, such as those on zombie
firms (Caballero, Hoshi, and Kashyap 2004, and Ahearne and Shinada 2004), have tended to focus on
the non-manufacturing sector, we need more research on why TFP growth slowed in the
manufacturing sector.
To sum up the above brief survey, the decline in TFP growth in Japan during the 1990s seems to
be more modest than suggested by Hayashi and Prescott. And although the majority of studies show
12
Cabinet Office (2002) and Hattori and Miyazaki (2000) obtain opposite results in their studies.
Based on growth accounting at the 2-digit industry level, these studies concluded that TFP growth in
the manufacturing sector did not substantially decline in 1990s. Moreover, the sharp decline of TFP
growth in the non-manufacturing sector contributed to the slowdown in macro TFP growth in the
1990s. Probably the following two factors are responsible for the different results. Firstly, the
Cabinet Office and Hattori and Miyazaki take account neither of changes in capacity utilization in
non-manufacturing sectors nor of changes in the quality of labor. Secondly, in order to evaluate each
factor’s contribution to output growth, the Cabinet Office study and the Hattori and Miyazaki study
use that factor’s distribution share, whereas Nishimura and Minetaki (2003) and Fukao et al. (2004)
use cost share. In the 1990s, the distribution share of labor was higher than the cost share of labor,
and labor input in the manufacturing sectors declined more drastically than in non-manufacturing
sectors. Because of this difference, the Cabinet Office study and the Hattori and Miyazaki study
arrive at a higher TFP growth in the manufacturing sector.
8
that there was some decline in TFP growth in Japan, we need more empirical research to purify the
Solow residuals and investigate the extent and nature of the slowdown in TFP growth. Another
important point is that many studies found that the slowdown in TFP growth in the manufacturing
sector was more severe than that in the non-manufacturing sector.
3. Decomposition Analysis of TFP Growth in the Manufacturing Sector
As Baily, Hulten and Campbell (1992) and Foster, Haltiwanger and Krizan (1998) have shown
in their productivity decomposition analyses, the start-up of productive establishments and the closure
of unproductive establishments substantially contributed to US TFP growth. As Figure 3.1 shows, the
start-up rate (number of newly opened establishments/number of all establishments) and the closure
rate in Japan are about one half of the corresponding values for the US in the 1980s. Moreover, the gap
widened in the 1990s. In particular, the start-up rate in Japan’s manufacturing sector has declined to
about 2% in recent years. Probably this factor has contributed to the slowdown in TFP growth in
Japan’s manufacturing sector. We examine these issues in this section.
We use the firm-level panel data underlying the Kigyo Katsudo Kihon Chosa (Basic Survey of
Japanese Business Structure and Activities) conducted annually by the Ministry of Economy, Trade
and Industry (METI).13 The survey covers all firms with at least 50 employees and 30 million yen of
paid-in capital in the Japanese manufacturing, mining and commerce sectors. We use the data for
manufacturing firms. Our data cover the period 1994–2001. After some screening of the data our
unbalanced panel data consists of 110,856 observations.14
13
The compilation of the micro-data of the METI survey was conducted as part of the project
“Foreign Direct Investment in Japan” at the Cabinet Office, Government of Japan.
14
We exclude all observations with zero values for material costs, compensation of employees, and
tangible fixed assets from our data set. We also exclude observations with an extremely high or low
capital-labor ratio. Through this screening process, the number of observations declined by about
9
Insert Figure 3.1
Following Good, Nadiri, and Sickles (1997) and Aw, Chen, and Roberts (1997), we define the
TFP level of firm f in year t in a certain industry in comparison with the TFP level of a hypothetical
representative firm in year 0 in that industry by15
n
1
ln TFPf ,t = (ln Y f ,t − ln Yt ) − ∑ ( S i , f ,t + S i ,t )(ln X i , f ,t − ln X i ,t )
i =1 2
t
t
n
1
+ ∑ (ln Ys − ln Ys −1 ) − ∑∑ ( S i , s + S i , s −1 )(ln X i , s − ln X i , s −1 )]
s =1
s =1 i =1 2
(2.1)
where Yf, t, Si, f, t, and Xi, f, t denote the gross output of firm f in year t, the cost share of factor i for firm
f in year t, and firm f’s input of factor i in year t, respectively. The variables with an upper bar denote
the industry average of that variable. We assume constant returns to scale. As factor inputs, we take
account of capital, labor and real intermediate inputs. For details on the definition and source of each
variable, please see Appendix A.16 Because of the limitation of the data we cannot take account of
the change in labor quality in our TFP analysis. It is probably because of this difference that we
arrive at a higher TFP growth estimate than the industry-level result in Fukao et al. (2004). We also
0.8% in comparison with our original set of observations.
15
We divide the manufacturing firm data into 30 sets of different industries and evaluated each
firm’s relative TFP level in relation to the industry average.
16
The approach used here also tries to deal with the following shortcomings of Nishimura et al.
(2003). First, Nishimura et al. used value-added instead of gross output as their output measure. As
Domar (1961) has shown, value-added-based TFP may differ from gross-output-based TFP, which is
commonly used in theoretical and empirical studies. Second, Nishimura et al. derived real
value-added using the value-added deflator of the SNA statistics. However, this deflator is based on
a relatively aggregated industry classification, so that their approach risks underestimating the TFP
growth of firms in high-tech industries, where output prices decline more rapidly. Compared with
their approach, we use the more disaggregated deflator of the Wholesale Price Statistics and
Corporate Goods Price Statistics.
10
assume that working hours and the capacity utilization rate at each firm are identical with those of
the industry average.
We define the representative firm for each industry as a hypothetical firm whose gross output as
well as input and cost share of all production factors are identical with the industry average. The first
two terms on the right hand side of equation (2.1) denote the gap between firm f’s TFP level in year t
and the representative firm’s TFP level in that year. The third and the fourth term denote the gap
between the representative firm’s TFP level in year t and the representative firm’s TFP level in year 0.
Therefore, lnTFPf, t in equation (2.1) denotes the gap between firm f’s TFP level in year t and the
representative firm’s TFP level in year 0.
Adopting the methodology used by Baily, Hulten and Campbell (1992) and Forster,
Haltiwanger and Krizan (1998), we define the industry-level TPF of a certain industry in year t by
n
ln TFPt = ∑θ f ,t ln TFPf ,t
(2.2)
f
where θf, t denotes firm f’s sales share in year t in that industry. Then, as Forster, Haltiwanger and
Krizan (1998) showed, we can approximate the manufacturing sector’s TFP growth from year τ to
year t, lnTFP t – lnTFPτ, by the sum of the following five factors.
Within effect:
∑θ
f ∈s
Between effect:
f , t −τ
∑ ∆θ
f ∈s
Covariance effect:
∑θ
f ∈N
Exit effect:
∑θ
f ∈X
f ,t
∑ ∆θ
f ∈s
Entry effect:
∆ ln TFPf ,t ,
f ,t
(ln TFPf ,t −τ − ln TFPt −τ ) ,
f ,t
∆ ln TFPf ,t ,
(ln TFPf ,t − ln TFPt −τ ) and
f , t −τ
(ln TFPt −τ − ln TFPf ,t −τ ) ,
where S is the set of firms that stayed in that industry from year τ to year t, N is the set of newly
11
entered firms and X is the set of exited firms.17 TFP with an upper bar denotes the industry-average
TFP level.
Our decomposition result for the period from 1994 to 2001 is reported in Table 3.1. It has been
pointed out in preceding studies that decomposition results are affected by business cycles.18 In order
to take this into account, we also conducted our decomposition on an annual basis. The results are
reported in Table 3.2. The switch-in and switch-out effects in Tables 3.1 and 3.2 denote the
contribution of those firms that moved from one industry to another to the industry average of the TFP
level. Table 3.3 compares our results with those of preceding studies for the US, South Korea, and a
number of European countries. We should note that our decomposition and the decomposition for the
European countries are based on firm-level data whereas the studies on the US and South Korea are
based on establishment-level data.
Insert Tables 3.1, 3.2, and 3.3
Our major findings are as follows.
1. Both the exit effect (excluding the switch-out effect) and the switch-out effect for the
manufacturing sector as a whole from 1996 to 2001 were negative and substantially contributed
to the decline in TFP growth in the manufacturing sector (Table 3.1). The negative exit effect
17
As already mentioned, the METI survey covers only those firms in the manufacturing and the
commerce sectors that are of a size that is greater than the cut-off level. Thus, our data on firms that
“exited” includes firms which shrunk or changed their main business from the manufacturing sector
to other sectors. We should also note that firms, which were merged and became part of another firm,
are treated as “exited.”
18
In 1990-2002, there were three official business cycle peaks, February 1991, May 1997, and
November 2000, and three troughs, October 1993, January 1999, and January 2002. Official peak
and trough dates are available in Business Cycle Reference Dates, Economic and Social Research
Institute, Cabinet Office, Government of Japan (<http://www.esri.cao.go.jp/>).
12
means that the average TFP level of exiting firms was higher than that of staying firms. Even
when we decompose TFP growth on an annual basis, we find that the exit effect (including the
switch-out effect) was negative for all seven years (Table 3.2). It is interesting to note that this
negative exit effect is not special to Japan. Italy and the Netherlands also experienced a negative
exit-effect (including the switch-out effect) in 1987-1992.
2. Both the entry effect (excluding the switch-in effect) and the switch-in effect were positive in
almost all the industries (Table 3.1). Moreover, the entry effect (including the switch-in effect)
was positive in both the upturn and the downturn periods (Table 3.2). But probably as a result of
the low entry rate, the size of the entry effect was not large. The entry effect (including the
switch-in effect) increased the TFP level of the manufacturing sector by 1.13% in 1994-2001,
which is much smaller than Korea’s entry effect in 1990-98 (15.60%) and Italy’s entry effect in
1987-1992 (5.12%) (Table 3.3).
3. The redistribution effect – that is, the between effect plus the covariance effect – was positive
(0.33%) but relatively small in comparison with that for the other countries (with the exception
of the Netherlands) (Table 3.3).
4. The within effect, i.e. the effect of TFP growth within staying firms, was the largest factor
among all the effects (Table 3.1). Moreover, this effect changed pro-cyclically (Table 3.2).
The above result suggests that in order to accelerate TFP growth in Japan’s manufacturing
sector it is important to promote new entries and to make both the exit process and the process of
resource allocation more efficient. These factors, moreover, are closely related with the allocation of
funds through the financial system. Therefore, the problems in Japan’s banking system are likely to
have contributed to the slowdown of Japan’s TFP growth, and their resolution forms an integral part
of any attempt to raise the TFP growth rate.
13
Unfortunately, the METI survey does not include detailed information on firms’ financial
affairs, such as each firm’s main bank or its interest payments for borrowing from banks. The only
available information is firms’ total liability. Using regression analysis based on pooled
cross-industry data, Fukao and Kwon (2003) found that there is a significant negative correlation
between the exit effect and that industry’s average liability-asset ratio.19 That is, in industries where
the liability-asset ratio is high, the exit effect tends to be negative. There is a possibility that the
malfunction of Japan’s financial system contributes to the negative exit effect by allowing zombie
firms to survive while high-productive small firms fail as a result of a credit crunch.
4. Inter-firm Differences in TFP
Following Japan’s recent economic recovery from the trough of January 2002, it has been
argued in newspapers and business journals that differences between firms in terms of their business
performances are increasing. In the case of the manufacturing sector, while large and internationalized
firms considerably managed to improve their performance, the performance of small and less
internationalized firms continued to stagnate.
Figure 4.1 compares the diffusion index (D.I.) of business conditions (“favorable” minus
“unfavorable”) for large and small manufacturing firms. The figure shows that the gap of the D.I.
between large and small firms has increased in recent years. While it is not unusual for this gap to
widen during a recovery, the recent extent is the largest in the past thirty years. Figure 4.2 compares
the labor productivity (in natural logarithm) of large and medium-sized firms on the one hand and
small firms on the other. Again, the gap has widened to an unprecedented level.
19
For this cross-industry regression, Fukao and Kwon (2003) divide the manufacturing firm data
into 58 sets of different industries and estimated the “exit effect” in each industry. Their result is also
reported in Fukao et al. (2004).
14
These trends suggest that there is a group of firms that has been excluded from recent
innovations and their stagnation hinders TFP growth in the manufacturing sector. In this section, we
examine this issue. Using the micro-data of the METI survey we measure the gap in the TFP level
between a group of high–TFP firms and a group of low–TFP firms and compare the characteristics of
these two groups.
Insert Figures 4.1 and 4.2
To date, the number of studies on the dispersion of productivity among Japanese firms is very
limited. Using the same firm-level data of the METI survey, Morikawa (2004) found that there is no
rising trend in the standard deviation of the current-profit/sales ratio or of the current-profit/total-asset
ratio of manufacturing firms in the period of 1991-2001. But he also found that the standard deviation
of sales growth substantially increased in 1991-2001.
Using firm-level data of publicly traded firms in the Development Bank of Japan (DBJ)
database, Shinada (2003) studied the technology gap between a group of firms at the technology
frontier and other firms in the manufacturing and the non-manufacturing sector for the period
1982-2000. He found that the technology gap widened in the 1990s when compared with the 1980s,
especially in the electrical machinery and equipment industry and in the automobile and auto-parts
industry. From the viewpoint of the TFP growth slowdown in the 1990s, Shinada’s finding is very
interesting.
A major shortcoming of Shinada’s analysis is that his dataset covers only a limited range of
activities in the Japanese economy. Compared with the number of all workers in each industry in 2001,
which we obtain from the Establishment and Enterprise Census for 2001 conducted by the Japanese
Ministry of Public Management, Home Affairs, Posts and Telecommunications, the total of the
15
number of workers employed by the firms covered by the DBJ Database in 2001 is very small. In the
case of the manufacturing sector, the DBJ Database covers 1,716 firms, accounting only for 24.7%
(=2,461,384/9,960,231) of the workforce overall. The coverage in the construction sector, in terms of
workers employed, is only 9.6%, while in the retail and wholesale sector (including eating and
drinking places) it is 5.9%, in the real estate sector it is 8.1%, and in the service sector it is 5.5%. The
METI survey covers a much larger number of firms: in the case of the manufacturing sector, the survey
covers 13,470 firms and 51.4% of the entire workforce. The coverage is still not complete, but much
better than in the case of the DBJ Database.
Table 4.1 shows how the TFP level gap (in natural logarithm) between the 75 percentile firm
and the 25 percentile firm changed over time in each industry in the period 1994-2001. We used the
same TFP data as in section 3. In the table, we placed industries in which the gap grew by a large
margin at the top of the table and those in which it declined by a large margin at the bottom. The
widening of the gap was particularly pronounced in drugs and medicine, petroleum and coal products,
and electronic data processing machines and electronic equipment. Since the gap widened in
large-sized industries, such as electronic data processing machines and automobiles, the average gap
of all the manufacturing industries, which is shown at the bottom of Table 4.1 also increased.20 In the
case of the average gap for the whole manufacturing sector, the widening occurred after 1997, the year
of Japan’s financial crisis.
According to preceding studies on other countries, the TFP gap moves counter-cyclically over
time (Bartelsman and Doms 2000). In 2001, the D.I. of business conditions for the whole
manufacturing sector was -41.8%, which is worse than the D.I. in 1994, -26.3. Therefore, there is some
risk that the TFP gap in 2001 is partly exaggerated by the recession. But even when we compare the
TFP gap in 2000, when the D.I. was as high as -11, with the TFP gap in 1994, we can observe an
20
To calculate the average value, we used each industry’s total sales as the weight.
16
increase in the TFP gap in many industries.21
Insert Table 4.1
Are there any common characteristics among the industries where the TFP gap widened? We
compared several industry characteristics which we expect to have a close relationship with the
productivity of firms. Figure 4.3 shows that the TFP gap widened mainly in industries with a high
R&D intensity and where the internationalization of firms is more advanced (we measure
internationalization by outward direct investment, the introduction of foreign capital, and
procurements from abroad).22, 23 There are statistically significant positive correlations between the
change of the TFP gap and three of the four characteristics, the R&D/sales ratio, the amount of
materials purchased from abroad/total amount of materials purchased, and the percentage of firms
owned by foreign firms. The correlation between the change in the TFP gap and the stock of direct
investment abroad/total assets was positive but statistically insignificant.
Insert Figure 4.3
21
In order to check the robustness of these results, we also calculated the change of the TFP gap for
the period 1995-2001. The results are displayed in Table 4.1 and are not qualitatively different.
22
As measures of industry characteristics, we used average values of pooled firm-data for each
industry.
23
The petroleum and coal products industry seems to have different characteristics from the other
industries where the TFP gap widened. In this industry, the R&D ratio and the stock of direct
investment abroad/total assets are very low. Probably we can explain the widening of the TFP gap in
this industry by the deregulation of 1997. Until 1997, imports of specific kinds of petroleum refined
products were regulated through a system of registration of importers, the Provisional Measures Law
on the Importation of Specific Kinds of Petroleum Refined Products. The gap between winners and
losers may have widened as a result of fierce competition after deregulation.
17
Next, we compare the characteristics, TFP growth, and the growth rate of sales and employment
of the top quartile firm-group and the bottom quartile firm-group. For each year and for each industry,
we select firms, whose TFP level is higher than the 75 percentile TFP level, as the top-quartile firms.
Panel A of Table 4.2 compares the pooled data of all the top-quartile firms with the pooled data of all
the bottom-quartile firms. The comparison between the top and the bottom firm-group shows that the
top firm-group is more internationalized than the bottom firm-group. The top-quartile firms show a
higher R&D intensity and the ratio of their number of non-production workers/number of all workers
is 11 percentage points higher than for the bottom-quartile firms. The top-quartile firms are also on
average 78% (= (exp (5.484-4.906)-1)*100) bigger (measured by the number of workers) and have a
liability/total asset ratio that is 16 percentage points lower. All these differences are statistically
significant.
Insert Table 4.2
It is also important to note that a large percentage of top-quartile firms is owned by other
domestic firms, and this percentage considerably increased after the currency crisis of 1997. In the
case of the pooled data of 1998-2001, 41 percent of top-quartile firms were majority-owned either by a
domestic or a foreign firm.
Table 4.2 also compares the performance of the top–quartile firm-group with the
bottom–quartile firm-group. The bottom firm-group exhibits a higher growth rate of both TFP and
sales. This phenomenon can probably be explained either by a convergence mechanism or by
temporary shocks. Table 4.2 also shows that the bottom-quartile firms reduce their employment more
rapidly and have a lower return-on-asset ratio.
18
Panel B of Table 4.2 compares the top-quartile firms with the bottom-quartile firms at the
industry level for six relatively large industries, where the TFP gap substantially widened. We can
confirm that almost all the above differences between the top and the bottom firm group hold in each
industry. Panel B also provides a number of further insights. The gaps in R&D intensity, the
percentage of foreign-owned firms, and in firm size between the top firms and the bottom firms are
largest in the drugs and medicine industry. The gap in the ratio of the stock of direct investment abroad
to total assets is largest in the automobile industry, while the gap in the percentage of
domestically-owned firms is largest in the electronic data processing machines and electronic
equipment industry. Finally, the gap in the number of non-production workers/number of all workers
is largest in the communication equipment and related products industry. These diverse patterns of
gaps among industries suggest that the main factor determining the advantage of the top firms is
different for different industries.
Our findings, so far, suggest the following tentative explanation for the recent widening of the
TFP gap among manufacturing firms. In the 1990s, many Japanese firms, especially large firms in
high-tech and globalized industries, further pressed ahead with internationalization and intensified
R&D efforts in order to improve their productivity. In such industries, the reorganization of
relationships among firms also proceeded through M&As. On the other hand, some firms, mainly
relatively small and borrowed-up firms, could not follow this innovation process and were left behind
in their productivity.
It is important to note that the causality behind these relationships could be the reverse. In other
words, the relationship may result not from the fact that characteristics such as a high R&D intensity
and a high degree of internationalization enhance firms’ productivity, but rather that only
high-productivity firms are able to conduct intensive R&D and internationalize and are targeted in
mergers and acquisitions. In order to examine this issue, we investigated the determinants of TFP
19
growth using our firm-level data.
Table 4.3 shows descriptive statistics of the variables used for the
regression analysis. The regression results are reported in Table 4.4.24
Insert Tables 4.3 and 4.4
We found that greater R&D intensity and internationalization have a positive effect on firms’
TFP growth. We also found that larger firms, firms owned by another domestic firm, firms with a
higher percentage of non-production workers in total workers and a lower liability-asset ratio have
higher TFP growth. These findings seem to support our hypothesis that a new divide caused by R&D,
internationalization, and reorganizations of relationships among firms through M&As is emerging and
growing in Japan’s manufacturing industry.
Our result also shows that firms with a lower TFP level tend to have higher TFP growth. As we
have already argued, we can explain this phenomenon either by a convergence mechanism or by
temporary shocks or noise. Some firms have lower TFP because of a temporary negative shock or
observation noise but the effect of the shock or the noise disappears in the next period.
In order to examine whether the determinants of TFP growth have changed in the estimation
period, we divided the period of 1995-2001 into two sub-periods, 1995-1997 and 1998-2001, and
estimated specification (4) of Table 4.4 for each sub-period. We did not find considerable changes in
the estimated coefficients except for a decline of the absolute value of the coefficient on the TFP level
(the beta-convergence coefficient). Running a regression for 1995-2001 with an additional cross term
of a dummy variable for 1998-2001 period and the TFP level in year t-1, we tested whether the there
24
Column 5 of Table 4.4 reports the results for the “long difference” estimator. In order to check the
robustness of our regression results based on annual TFP growth, we regressed the TFP growth from
1994 to 2001 on the variables for 1994. The results we obtained were qualitatively similar to those
based on annual TFP growth.
20
was a decline in the absolute value of the beta-convergence coefficient. The result is reported in Table
4.5. We found that the absolute value of the negative beta-convergence coefficient declined
significantly in the later sub-period. In the case of specifications (A) and (B) in Table 4.5, the
coefficient on beta-convergence declined from -0.386 to -0.235. These values mean that the half-life
of the GDP gap increased from 1.42 years (=ln(0.5)/ln(1-0.386)) in 1995-1997 to 2.59 years
(=ln(0.5)/ln(1-0.235)) in 1998-2001. We also ran this regression for each industry and found that in
many industries, the absolute value of the beta-coefficient declined.
Insert Table 4.5
The slowdown of the beta-convergence mechanism seems to have contributed to the widening
of the TFP dispersion among firms. Until the middle of the 1990s, large firms and many small firms
were closely tied by sub-contracting and keiretsu relationships and it seems that through this network
advanced technologies of assemblers and key-component producers were transferred to lower-tier
small suppliers. 25 But as large firms relocated their production abroad and rationalized their
procurement processes, this technology-transfer mechanism probably slowed down.
According to our interview, Toshiba now makes almost all decisions on procurements of parts
and components for all of its three notebook PC factories at its headquarters in Tokyo taking a global
viewpoint. Toshiba’s PC factories are located in Hangzhou, China, the Laguna Technopark in the
Philippines and in Ome, near Tokyo. In the case of Toshiba Hangzhou, the major suppliers are
Japanese, Taiwanese, and Korean affiliates in China and firms located in East Asia. Another good
25
Using plant-level data for Japan’s automobile industry for 1981-1996, Ito (2002) found that parts
producers, which were located close to assembler’s plant, tended to achieve higher TFP growth than
other parts producers.
21
example is Nissan Motors. Based on its “Revival Plan” Nissan Motors drastically reduced the number
of its suppliers and introduced more competition into its procurement process.
So far, we found that compared with firms with low TFP, firms with high TFP tend to be more
R&D intensive and more internationalized. High-TFP firms also typically are larger and have a lower
liability/total asset ratio. These differences further raise the TFP growth rate of the firms with a high
TFP level. On the other hand, we also found a mechanism of convergence, which is probably caused
by technology spillovers and catching up. From a quantitative viewpoint, how much do these two
factors contribute to the dispersion of TFP among firms? In order to answer this question we need a
model.26 Let us assume that specification (4) of Table 4.4 is the correct model and the dynamics of
firms’ TFP level are determined by this equation.27 Then, the larger the dispersion of firms’
26
On theories on the dispersion of firms’ TFP level, see Baily et al. (1992) and Bartelsman and
Doms (2000).
27
If we assume that all firm characteristics except the TFP level are exogenously determined and
constant, then the unconditional standard deviation of the TFP level among firms depends on three
factors. Suppose that firm i’s TFP level (in natural logarithm), yi,t is determined by the following
dynamics.
n
yi ,t +1 − yi ,t = − β yi ,t + ∑ γ i xi + ui ,t
i =1
where xj,i denotes firm i’s j’th characteristic. ui,t denotes a random shock. We assume that β satisfies
0<β<1. Then we have
∞
n

 ∞
yi ,t +1 = ∑ (1 − β )τ ∑ γ i xi  + ∑ (1 − β )τ u i ,t −τ .
i =1
τ =0 
 τ =0
The unconditional standard deviation of the TFP level among firms is given by
σ ( y) =
n
1

 σ (∑ γ i xi ) + σ (u ) 
β  i =1

where σ denotes the standard deviation of each variable.
Unfortunately, it seems unrealistic to assume that firms’ characteristics except the TFP level are
exogenously determined and constant. For this reason, we did not calculate the unconditional
standard deviation of the TFP level.
22
characteristics, the larger will be the dispersion of the TFP level. Similarly the larger the standard
deviation of the random shock or the smaller the absolute value of the beta-convergence coefficient,
the larger will be the unconditional standard deviation of the TFP level.
In Table 4.6 we compared the size of the random shock, the effect of the beta convergence and
the divergence effect caused by firms’ characteristics. Table 4.6 shows that the disparity of firms’
characteristics between the top quartile firm-group and the bottom quartile firm-group continuously
worked to widen the TFP gap between the two groups by about 1% a year. On the other hand, the
beta-convergence mechanism became weak in the period 1998-2001. There was no considerable
change in the size of random shocks.
Insert Table 4.6
Next we show the degree of persistence of firms’ relative productivity level among continuing
firms. Table 4.7 shows the transition matrix. According to this table, the degree of persistence is very
high. More than one-half of firms which originally ranked in the bottom three deciles in 1994
remained in the same three deciles in 2001. Similarly more than one-half of firms which originally
ranked in the top three deciles in 1994 remained in the same three deciles in 2001.28
Insert Table 4.7
If the bottom firms reduce employment or exit and the top firms stay and expand employment,
the macro-level TFP will increase. To conclude this section, we analyze this “metabolism” issue.
28
In the case of plant-level TFP in the US manufacturing sector, several studies have shown that the
degree of persistence is very high (Baily et. al 1992, Bartelsman and Doms 2000).
23
Table 4.8 compares firms’ employment growth and firm “exits” between the top firm-group and the
bottom firm-group. The table shows that the reduction of the number of workers by the low–TFP
firms is not much larger than the reduction by the high–TFP firms. Thus, while we might expect that
high-TFP firms would be expanding employment and output, this is not the case. The likely reasons
are that high-TFP firms are in the process of restructuring their activities and many of them are
relocating production abroad. This finding suggests that the “metabolism” – the expansion of
employment and output by high-TFP firms and the contraction or exit of low-TFP firms – is not
working well in Japan’s manufacturing sector.
Insert Table 4.8
5. Conclusion
Our results can be summarized as follows. Using firm-level data of the METI survey, we
examined why Japan’s TFP growth slowed down in the manufacturing sector. Our decomposition
analysis showed that the exit effect was negative and substantially contributed to the decline in TFP
growth in the manufacturing sector. The negative exit effect means that the average TFP level of
exiting firms was higher than that of staying firms. We also found that although both the entry effect
and the redistribution effect were positive, they were very small when compared with those in other
countries. This “low metabolism” seems to have slowed down the TFP growth of the manufacturing
sector.
In section 4 we measured the gap of the TFP level between a group of high–TFP firms and a
group of low–TFP firms and compared the characteristics of these two groups. We found that the TFP
gap between the 75 percentile firm and the 25 percentile firm is widening in many industries where
R&D intensity is high and the internationalization of firms is more advanced. The TFP gap seems to
24
be widening because high–TFP firms tend to have a higher R&D intensity, a higher degree of
internationalization, are larger, and have a lower liability-asset ratio, and these characteristics enhance
their productivity further. It is also important to note that a large percentage of top–quartile firms are
owned by other domestic firms. In the case of the pooled data for 1998-2001, 41 percent of top quartile
firms were majority owned either by a domestic a foreign firm. We also found that the catching-up
mechanism of low-productive firms slowed down after 1997.
25
Appendix A: Definition of Variables Used in the Econometric Analysis and Data Sources
We used each firm’s total sales and cost of intermediate inputs as nominal gross output and
nominal intermediate input data. We derived the deflator for each industry’s gross output and
intermediate input from the Bank of Japan’s Wholesale Price Statistics and Corporate Goods Price
Statistics.
For capital stock, the only data available are the nominal book values of tangible fixed assets in
the Basic Survey of Japanese Business Structure and Activities. Using these data, we calculated the net
capital stock of firm f in industry j in constant 1995 prices as follows:
K ft = BV ft ∗ ( INK jt / IBV jt )
where BVft represents the book value of firm f’s tangible fixed capital in year t, INKjt stands for the
net capital stock of industry j in constant 1995 prices, and IBVjt denotes the book value of industry
j’s capital. INKjt is calculated as follows. First, as a benchmark, we took the data on the book value
of tangible fixed assets of year 1976 from the Census of Manufactures 1976 published by METI. We
then converted the book value of year 1976 into the real value in constant 1995 prices using the net
fixed assets deflator provided in the Annual Report on National Accounts published by the Cabinet
Office, Government of Japan. Second, the net capital stock of industry j, INKjt, for succeeding years
was calculated using the perpetual inventory method. We used the capital formation deflator in the
Annual Report on National Accounts and Masuda’s (2000) estimate of the depreciation rate of
0.0792 for the calculation.
In order to obtain capital input, we multiplied the net capital stock by the capital utilization ratio
of each industry provided in the JIP database.29
29
The JIP Database was compiled as part of an ESRI (Economic and Social Research Institute,
Cabinet Office, Government of Japan) research project. The detailed result of this project is reported
in Fukao, Miyagawa, Kawai, Inui (2004). The database contains annual information on 84 sectors,
including 49 non-manufacturing sectors, from 1970 to 1998. These sectors cover the whole Japanese
26
As labor input, we used each firm’s total number of workers multiplied by the sectoral
working-hours from the Ministry of Health, Labour and Welfare’s Monthly Labor Survey. We were not
able to take account of differences in labor quality among firms, though it seems fair to assume that a
group of high-TFP firms probably tend to employ more educated workers. Our estimates of TFP level
might be biased upwards for high-TFP firms as a result of this neglect of the labor quality.
Finally, we derived the cost shares of the factors of production. For labor cost, we used the wage
data provided in the Basic Survey of Japanese Business Structure and Activities. Intermediate input
cost is defined as total production cost plus cost of sales and general management minus wages minus
depreciation. Capital cost was calculated by multiplying the real net capital stock with the user cost of
capital. The latter was estimated as follows:
1 − τz
dq
Pk = q ∗ (
)[r + δ k − ]
1−τ
q
where q, r , δ ,τ and z are the prices of investment goods, interest rates, depreciation rates,
corporate tax rates, and the present values of depreciation deduction on a unit of nominal investment,
respectively. Data on investment goods prices, interest rates, and corporate tax rates were taken from
the Annual Report on National Accounts, the Ministry of Finance Statistics Monthly. The
depreciation rate for each industry is estimated using the book value of tangible fixed assets at the
beginning of year t and the depreciation expense during year t in the Census of Manufactures
published by METI.
economy. The database includes detailed information on factor inputs, annual nominal and real
input-output tables, and some additional statistics, such as R&D stock, capacity utilization rate,
Japan’s international trade statistics by trade partner, inward and outward FDI, etc., at the detailed
sectoral level. An Excel file version (in Japanese) of the JIP Database is available on ESRI’s web
site.
27
References
Ahearne, Alan G., and Naoki Shinada (2004) “Zombie Firms and Economic Stagnation in Japan,”
paper presented at the University of Michigan CGP Conference, Macro/Financial Issues and
International Economic Relations: Policy Options for Japan and the United States, October
22-23, 2004.
Ahn, Sanghoon, Hyeog Ug Kwon, and Kyoji Fukao (2004) “The Internationalization and
Performance of Korean and Japanese Firms: An Empirical Analysis Based on Micro Data”
mimeo, Hitotsubashi University.
Aw, Bee Yan, Xiaomin Chen, and Mark J. Roberts (2001) “Firm-level Evidence on Productivity
Differentials and Turnover in Taiwanese Manufacturing,” Journal of Development Economics,
vol. 66, no.1, pp. 51-86.
Baily, Martin Neil, Charles Hulten, and David Campbell (1992) “Productivity Dynamics in
Manufacturing Plants,” Brookings Papers on Economics Activity: Microeconomics, vol. 2, pp.
187-249.
Barnes, Matthew, Jonathan Haskell, and Mika Maliranta (2001) “The Sources of Productivity
Growth: Micro-Level Evidence for the OECD,” paper presented at the OECD Workshop on
Firm-Level Statistics, November 26-27, 2001.
Bartelsman, Eric J. and Mark Doms (2000) “Understanding Productivity: Lessons from Longitudinal
Microdata,” Journal of Economic Literature, vol. 38, pp. 569-594.
Basu, Susanto and Miles S. Kimball (1997) “Cyclical Productivity with Unobserved Input
Variation,” NBER Working Paper, no. 5915.
Caballero, Ricardo J., Takeo Hoshi, and Anil K Kashyap (2004) “Zombie Lending and Depressed
Restructuring in Japan,” paper presented at Hitotsubashi University, Macro-Money Workshop,
June 18th, 2004.
Cabinet Office, Government of Japan (2002), Annual Report on the Japanese Economy and Public
Finance 2001-2002: No Gains without Reforms II, Cabinet Office, Government of Japan,
Tokyo.
Colecchia, Alessandra, and Paul Schreyer (2002) “ICT Investment and Economic Growth in the
1990s: Is the United States a Unique Case?,” Review of Economic Dynamics, vol. 5, pp.
408-422.
Domar, Evsey (1961) “On the Measurement of Technological Change,” Economic Journal, vol. 71,
no. 284, pp. 709-29.
Dunne, Timothy, Lucia Foster, John Haltiwanger, and Kenneth Troske (2000) “Wage and
Productivity Dispersion in U.S. Manufacturing: The Role of Computer Investment,” NBER
Working paper, no.7465.
Foster, Lucia, John Haltiwanger, and C. J. Krizan (1998) “Aggregate Productivity Growth: Lessons
28
from Microeconomic Evidence,” NBER Working Paper, no. 6803.
Fukao, Kyoji, and Hyeog Ug Kwon (2003) “Nippon no Seisansei to Keizai Seicho (The Productivity
and Economic Growth of Japan),” paper presented at the Semi-annual Conference of the Japan
Economic Association, June 14th, 2003, Oita.
Fukao, Kyoji, and Hyeog Ug Kwon (2004) “Nippon no Seisansei to Keizai Seicho: Sangyo Reberu
Kigyo Reberu Deta niyoru Jissho Bunseki (The Productivity and Economic Growth of Japan:
Empirical Analysis based on Industry-Level and Firm-Level Data),” Keizai Kenkyu, vol. 55,
no. 3, pp. 261-284.
Fukao, Kyoji, Tomohiko Inui, Hiroki Kawai, and Tsutomu Miyagawa (2004) "Sectoral Productivity
and Economic Growth in Japan, 1970-98; An Empirical Analysis Based on the JIP Database,"
in Takatoshi Ito and Andrew K. Rose, eds., Growth and Productivity in East Asia, National
Bureau of Economic Research-East Asia Seminar on Economics, vol. 13, Chicago, IL:
University of Chicago Press..
Fukao, Kyoji, Tsutomu Miyagawa, Hiroki Kawai, and Tomohiko Inui (2003) "Sangyo Betsu
Seisansei to Keizai Seicho: 1970-98 (Sectoral Productivity and Economic Growth: 1970-98),"
Keizai Bunseki, no. 170, Economic and Social Research Institute, Cabinet Office, Government
of Japan, Tokyo.
Fukao, Mitsuhiro (2003) “Choki Fukyo no Shuin wa Juyo Fusoku ni Aru (The Major Cause of the
Long Recession is the Lack of Demand),” in Kikuo Iwata and Tsutomu Miyagawa, eds.,
Ushinawareta Junen no Shinin wa Nanika (What is the Real Cause of the Lost Decade, Toyo
Keizai Shinpo Sha.
Good, David, H., M. Ishaq Nadiri, and Robin C. Sickles (1997) “Index Number and Factor Demand
Approaches to the Estimation of Productivity,” Handbook of Applied Econometrics vol. 2:
Microeconometrics, pp. 14-80.
Griliches, Zvi, and Haim Regev (1995) “Productivity and Firm Turnover in Israeli Industry:
1979-1988,” Journal of Econometrics, vol. 65, no.1, pp.175-203.
Hattori, Tsuneaki, and Miyazaki Hironobu (2000) “Sangyo Betsu no Gijutsu Shinporitsu no Keisoku
to Keizai Seicho no Yoinbunkai – 1970 Nendai Kohan Iko no Jisshokenkyu (Measuring
Technological Progress and Factor Analysis of Economic Growth in Japanese Industry),”
Denryoku Keizai Kenkyu, vol.44, pp.1-16.
Hayashi, Fumio, and Edward C. Prescott (2002) “The 1990s in Japan: A Lost Decade,” Review of
Economic Dynamics, vol. 5, no. 1, pp. 206-35.
Inui, Tomohiko and Hyeog Ug Kwon (2004) “Tembo: Nihon no TFP Joshoritsu wa 1990-nendaini
oite Doredake Teika Shitaka (Survey: Did the TFP Growth Rate in Japan Decline in the
1990s),” ESRI Discussion Papers Series, no. 115.
Ito, Keiko (2002) “Plant Productivity, Keiretsu, and Agglomeration in the Japanese Automobile
29
Industry: An Empirical Analysis Based on Micro-Data of Census of Manufactures
1981-1996,” mimeo, Senshu University.
Jorgenson, Dale W., and Kazuyuki Motohashi (2003) “The Role of Information Technology in the
Economy: Comparison between Japan and the United States,” prepared for RIETI/KEIO
Conference on Japanese Economy: Leading East Asia in the 21st Century? Keio University,
May 30, 2003.
Kawamoto, Takiji (2004) “What Do the Purified Solow Residuals Tell Us about Japan’s Lost
Decade?” Bank of Japan IMES Discussion Paper Series, no. 2004-E-5, Bank of Japan, Tokyo.
Masuda, Muneto (2000) “Shihon Shutokku Tokei no Mikata: Shijo Hyoka Shihon Sutokku no
Shisan (A Perception of Japan’s Capital Stock Statistics: A Trial Calculation of Capital Stock
in Market Value),” Research and Statistics Department Discussion Paper Series, No.00-5,
Bank of Japan, Tokyo.
Miyagawa, Tsutomu (2003) “Ushinawareta Junen to Sangyo Kozo no Tenkan (The Lost Decade and
Structural Change),” in Kikuo Iwata and Tsutomu Miyagawa eds., Ushinawareta Junen no
Shinin wa Nanika (What is the Real Cause of the Lost Decade, Toyo Keizai Shinpo Sha.
Morikawa, Masayuki (2004) “Are Japanese Firms ‘Polarizing’ in their Performances? – An
Empirical Test Based on the Micro-Data of the Basic Survey on Business Activities by
Enterprises – (Nihon Kigyo no Gyoseki wa ‘Nikyokuka’ Shite Iru ka? – Kigyo Katsudo Kihon
Chosa Maikuro Deita niyoru Kensho –),” Chosa Working Paper, no. WP04-05, Research
Section, Economic and Industrial Policy Bureau, Ministry of Economy, Trade and Industry,
Government of Japan, Tokyo.
Morrison, Catherine J. (1993) A Microeconomic Approach to the Measurement of Economic
Performance: Productivity Growth, Capacity Utilization, and Related Performance Indicators,
Springer-Verlag, New York.
New Business Creation Subcommittee, New Growth Policy Committee, The Industrial Structure
Council (2002) New Business Creation Subcommittee Report: For the Promotion of New
Business and the Facilitation of Growth, Ministry of Economy, Trade and Industry,
Government of Japan.
Nishimura, Kiyohiko, G., Takanobu Nakajima, and Kozo Kiyota (2003) “Does the Natural Selection
Mechanism Still Work in Severe Recessions? Examination of the Japanese Economy in the
1990s.” Journal of Economic Behavior and Organization, forthcoming.
Nishimura, Kiyohiko, and G., Kazunori Minetaki (2003) Joho Gijutsu Kakushin to Nihon Keizai
(Innovation in Information Technology and the Japanese Economy), Yuhikaku, Tokyo.
Shinada, Naoki (2003) “Decline in Productivity in Japan and Disparities between Firms in the
1990s: An Empirical Approach Based on Data Envelopment Analysis,” Development Bank of
Japan Research Report, No. 38, Development Bank of Japan, Tokyo.
30
Small and Medium Enterprise Agency, Ministry of Industry, Trade and Industry, Japanese
Government (2001) 2001 White Paper on Small and Medium Enterprises in Japan, Ministry
of
Industry,
Trade
and
Industry,
Japanese
Government,
Tokyo
(available
at
http://www.chusho.meti.go.jp /sme_english/index.html).
Small Business Administration, US Government (1998) The State of Small Business: A Report of the
President, Small Business Administration, US Government, Washington D.C.
Study Group on “Industry Hollowing-out” and Tariff Policy, Ministry of Finance, Japanese
Government (2002) Chairperson’s Report, Ministry of Finance, Japanese Government, Tokyo.
Yoshikawa, Hiroshi (2003) “Hayashi Ronbun eno Komento: Sugitaru wa Nao Oyobazaru ga
Gotoshi !? (Comment on Hayashi Paper: Too Much is as Bad as Too Little !?)” in Kikuo Iwata
and Tsutomu Miyagawa eds., Ushinawareta Junen no Shinin wa Nanika (What is the Real
Cause of the Lost Decade, Toyo Keizai Shinpo Sha.
Yoshikawa, Hiroshi and Kazuyuki Matsumoto (2001), 1990-nendai no Nichibei Keizai (The
Japanese and the US Economy in the 1990s), Financial Review, vol. 58, Policy Research
Institute, Ministry of Finance, Government of Japan, Tokyo.
31
Table 2.1 Comparison of Empirical Studies on Japan's TFP growth in 1990s
Capital services
Study
Period
Outputs
Labor Services
Share
Capital stock and capital service
prices
Capital utilization
Labor quality
Hours worked
Productio
Adjustment of
Assumptions on
n
deflator for IT
market structure technolog
products
y
Hayashi and Prescott
(2002)
1960-2000
GNP (based on
1968 SNA)
Net capital stock estimated by
the perpetual inventory method.
Net foreign assets and inventories
are included in the capital stock.
Unadjusted
Unadjusted
Adjusted by total
hours worked
Income share
within value
added (fixed)
Unadjusted
Perfect
competition in
both output and
input markets
Yoshikawa and
Matsumoto (2001)
1980-1999
GDP (based on
1968 SNA),
sectoral value
added
Gross capital stock of private
enterprises (statistics published
by the Cabinet Office)
Unadjusted
Unadjusted
Not reported
Income share
within value
added (fixed)
Unadjusted
Perfect
competition in
both output and
input markets
CRS
GDP (based on
1968 SNA),
sectoral value
added
Adjusted by capacityGross capital stock of private
enterprises (statistics published utilization rate only in the case
of manufacturing sectors
by the Cabinet Office)
Unadjusted
Income share
within value
added (fixed)
Unadjusted
Perfect
competition in
both output and
input markets
CRS
GDP (based on
1993 SNA),
sectoral value
added
Gross capital stock of private
enterprises (statistics published
by the Cabinet Office)
Adjusted by capacityutilization rate both in the
manufacturing sectors and in
the non-manufacturing sectors
Unadjusted
Adjusted by hours Income share
worked in each
within value
sector
added (fixed)
Unadjusted
Perfect
competition in
both output and
input markets
CRS
Sectoral value
added (based on
1968 SNA)
Net capital stock and service
prices of five capital goods
(JCER database)
Adjusted by assuming that the
capital is a quasi-fixed factor
Adjusted by IT
deflator of the
US
("harmonized
approach")
Perfect
competition in
input markets
CRS
GDP (based on
1993 SNA),
sectoral value
added
Net capital stock and service
prices of five capital goods
(JCER database)
Adjusted by capacityutilization rate in the case of
manufacturing sectors and
adjusted by information on
electricity input in the case of
non-manufacturing sectors
Unadjusted
Adjusted by hours Income share
worked in each
within value
sector
added (fixed)
Unadjusted
Perfect
competition in
both output and
input markets
CRS
GDP (based on
1968 SNA),
sectoral gross
output
Net capital stock and service
prices of 37 capital goods (JIP
database)
Adjusted by capacityutilization rate in the case of
manufacturing sectors and
adjusted by diffusion indices
on excess capacity (BOJ) in
the case of non-manufacturing
sectors (JIP Database)
Adjusted by
gender, age and
education
attainment (JIP
database)
Adjusted by hours
worked in each
sector
Unadjusted
Perfect
competition in
input markets
CRS
Adjusted by
gender, age,
education
attainment, and
type of occupation
（KEO database)
Adjusted (KEO
database)
Perfect
competition in
input markets
CRS
Adjusted by
gender, age and
education
attainment (JIP
database)
Adjusted by hours
worked in each
sector (statistics
published by the
Ministry of
Welfare and
Labor)
Hattori and Miyazaki
1978-1997
(2000)
Cabinet Office,
Government of Japan 1981-2000
(2002)
Nishimura and
Minetaki (2003)
Miyagawa (2003)
1975-1998
1981-1999
Fukao, Inui, Kawai,
1970-1998
and Miyagawa (2003)
Jorgenson and
Motohashi (2003)
Kawamoto (2004)
Based on their own
1975-1998
definition of GDP
1973-1998
GDP (based on
1968 SNA),
sectoral gross
output
Source: Inui and Kwon (2004) and the papers listed.
Net capital stock and service
prices of 62 capital goods plus
input of land, inventory stock,
software, and durable
consumption goods
Not explicitly adjusted
Net capital stock and service
prices of 37 capital goods (JIP
database)
Changes in hours per worker
is used as a proxy for
unobserved changes in both
labor effort and capital
utilization
Adjusted by total
hours worked
Adjusted by age,
Adjusted by hours
education
worked in each
attainment, and
sector
type of occupation
Cost share
Cost share
Cost share
Cost share
Adjusted by IT
deflator of the
US
("harmonized
approach")
Unadjusted
Perfect
competition in
input markets
Estimated annual TFP growth rate
Manufacturing
sector
Nonmanufacturing
sector
1980-90: 1.20%
1980-90: 2.5%
1980-90: 0.5%
1990-98: -0.90%
1990-98: 0.3%
1990-98: -1.3%
Macro level
Constant
1983-91: 2.36%
returns to
scale
1991-00: 0.19%
(CRS)
1987-93: 1.20% 1987-93: 1.61% 1987-93: 1.61%
1994-97: -0.60% 1994-97: 1.54% 1994-97: -1.43%
1981-90: 1.60%
1981-90: 2.1%
1981-90: 1.3%
1991-00: 0.20%
1991-00: 1.8%
1991-00: -0.3%
1981-89: 2.67% 1981-89: 1.21%
1990-98: 0.30% 1990-98: -0.26%
1981-90: 1.63% 1981-90: 2.81% 1981-90: 1.18%
1991-99: 0.84% 1991-99: 1.37% 1991-99: 0.64%
1983-91: 0.40% 1983-91: 0.78% 1983-91: -0.15%
1991-98: 0.03% 1991-98: -0.16% 1991-98: 0.27%
1975-90: 1.01%
1990-98: 0.89%
Private sector
1980-90: 1.9
CRS is not
assumed.
Private sector
1990-98: 1.9
Durable
manufacturing
1980-90: 2.8%
1990-98: 1.4%
Non-durable
manufacturing
1980-90: 1.7%
1990-98: 1.9%
1980-90: 1.6%
1990-98: 2.1%
Figure 3.1 Start-up and Closure Rate of Establishments: Japan-US Comparison
Figure 4.2.Panel B. Closure Rate:
Japan-US Comparison %
Figure 4.1.Panel A. Start-up Rate:
Japan-US Comparison %
16
16
All industries: US
12
12
10
10
8
8
6
6
4
4
2
2
0
0
All industries:
Japan
Wholesale, retail
and restaurants:
Japan
Services: Japan
Manufacturing:
Japan
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
14
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
14
Both the US and the Japanese data are based on statistics of employment insurance program.
Sources: Small Business Administration, US Government (1998), Small and Medium Enterprise Agency, Ministry of
Industry, Trade and Industry, Japanese Government (2001), and Study Group on “Industry Hollowing-out” and Tariff
Policy, Ministry of Finance, Jap
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Table 3.1 Decomposition of Sectoral TFP Growth: 1994-2001 (Growth over the Seven Year Period)
Exit effect
Entry effect
Total effect
(excluding
Switch-in
(excluding
Covariance
Industry
among
Within effect Between effect
switch-out
effect
switch-in
effect
stayers
effect)
effect)
b
c
e
f
g
a
d=a+b+c
Food
-0.001
0.001
0.006
-0.005
0.001
0.003
0.003
Textiles
-0.002
0.007
0.007
-0.027
0.001
-0.008
-0.002
Wood and furniture
-0.002
0.001
0.004
-0.017
0.001
0.000
0.000
Pulp and paper
0.000
0.001
0.000
-0.002
0.000
0.007
0.008
Printing and publishing
-0.001
0.003
0.010
0.000
0.000
0.008
0.011
Industrial chemicals and chemical fibers
0.001
0.002
0.001
-0.001
0.002
0.007
0.010
Oils and paints
-0.001
0.001
0.003
-0.004
0.001
0.004
0.004
Drugs and medicine
0.005
0.022
0.010
0.000
0.001
0.073
0.100
Other chemical products
0.001
0.000
0.007
0.001
0.002
0.016
0.017
Petroleum and coal products
0.015
0.014
0.000
-0.010
0.000
0.027
0.056
Plastic products
-0.002
0.003
0.004
-0.002
0.005
-0.003
-0.001
Rubber products
0.002
0.000
0.000
0.000
0.001
-0.024
-0.022
Ceramics
-0.003
0.003
0.002
-0.004
0.003
0.013
0.013
Iron and steel
-0.005
0.000
0.000
0.001
0.001
0.003
-0.002
Non-ferrous metals and products
-0.001
0.002
-0.006
-0.005
0.001
0.008
0.009
Fabricated metal products
-0.006
0.006
0.003
-0.003
0.003
-0.009
-0.009
Metal working machinery
-0.002
0.015
0.009
0.000
0.012
0.037
0.050
Special industrial machinery
-0.003
-0.002
0.003
-0.004
0.003
0.022
0.017
Office, service industry and household machines
-0.001
0.009
0.013
0.000
0.045
0.040
0.048
Miscellaneous machinery and machine parts
-0.002
0.002
0.003
-0.003
0.002
-0.005
-0.005
Industrial electric apparatus
-0.005
0.002
0.005
-0.003
0.001
-0.015
-0.018
Household electric appliances
-0.002
-0.002
0.007
0.001
0.005
0.015
0.010
Communication equipment and related products
-0.001
0.010
0.008
0.003
0.108
0.071
0.081
Electronic data processing machines and electronic equipment
0.003
0.005
0.008
-0.002
0.004
0.026
0.033
Electronic parts and devices
-0.001
0.010
0.017
-0.008
0.007
0.044
0.053
Miscellaneous electrical machinery and supplies
0.000
0.004
0.033
-0.012
0.011
0.014
0.017
Motor vehicles
0.000
0.005
0.001
-0.001
0.000
0.018
0.023
Miscellaneous transportation equipment
0.000
0.002
0.005
0.002
0.003
0.040
0.041
Precision instruments
-0.002
0.011
0.011
0.003
0.003
0.017
0.026
Other manufacturing
0.004
-0.001
0.046
-0.005
0.009
-0.005
-0.002
Weighted average of all the industries
-0.001
0.004
0.006
-0.004
0.005
0.012
0.015
Share of each effect in industry's TFP growth
-0.04
0.20
0.28
-0.17
0.25
0.56
0.71
Switch-out
effect
Net-entry
effect
h
i=e+f+g+h
0.002
-0.019
-0.027
-0.001
0.010
0.000
0.000
0.010
0.012
-0.010
0.006
0.000
0.009
0.000
-0.011
0.000
0.015
0.000
0.055
-0.001
0.000
0.010
0.120
-0.002
0.014
0.025
0.000
0.011
0.016
0.046
0.006
0.29
0.000
0.000
-0.015
0.000
0.000
0.000
0.000
0.000
0.001
0.000
-0.001
-0.001
0.008
-0.001
-0.001
-0.003
-0.007
-0.001
-0.002
-0.002
-0.003
-0.002
0.002
-0.012
-0.002
-0.007
0.000
0.000
-0.001
-0.004
-0.001
-0.07
Industry total
j=d+i
0.005
-0.020
-0.028
0.007
0.020
0.011
0.005
0.110
0.028
0.046
0.005
-0.022
0.022
-0.002
-0.002
-0.009
0.065
0.018
0.103
-0.006
-0.018
0.020
0.201
0.031
0.067
0.042
0.022
0.052
0.041
0.045
0.021
1.00
Table 3.2 Decomposition of Annual TFP Growth in the Manufacturing Sector
Contribution of each effect
Exit effect
Entry
TFP growth
Redistribu
Net entry
Between Covariance
(excluding Switch-in Switch-out
effect
Within
Period
total
tion effect
effect
effect
effect
effect
effect
(excluding switch-out
effect
subtotal
subtotal
switch-in
effect)
g
h
i
j
d
e
a=b+c+f
b
c=d+e
f=g+h
-0.002
0.002
0.006
-0.003
0.006
-0.003
1994-1995
0.029
0.024
0.000
0.005
0.000
0.002
0.004
-0.002
0.005
-0.005
1995-1996
0.011
0.008
0.001
0.002
0.001
0.002
0.001
-0.004
0.004
-0.004
1996-1997
-0.002
-0.002
0.003
-0.003
0.000
0.002
0.003
-0.003
0.002
-0.001
1997-1998
-0.007
-0.008
0.001
0.000
-0.002
0.002
0.003
-0.003
0.002
-0.002
1998-1999
0.011
0.010
0.000
0.001
0.001
0.002
0.004
-0.004
0.002
-0.001
1999-2000
0.017
0.013
0.003
0.001
-0.001
0.004
0.003
-0.004
0.003
-0.003
2000-2001
-0.005
-0.008
0.003
-0.001
Note: The entry and exit effect in this paper include the switch-in and switch-out effect, respectively.
Table 3.3 Comparison of Total Factor Productivity Decompositions of Each Country's Manufacturing Sector Based on Foster, Haltiwanger, and Krizan Method
Contribution of each effect
TFP
Entry effect Exit effect
Redistributio
Net entry
growth
Within
Between Covariance
(including (including
Source
Country
Unit of analysis
Period
n effect
effect
total (%)
effect
effect
effect
switch-in switch-out
subtotal
subtotal
effect)
effect)
a=b+c+f
b
c=d+e
d
e
f=g+h
g
h
Ahn, Kwon, Fukao (2004)
Korea
Establishment
1990-98
28.1
11.35
(0.40)
0.63
(0.02)
-2.28
(-0.08)
2.90
(0.10)
16.11
(0.57)
Foster, Haltiwanger, and Krizan (1998)
USA
Establishment
1977-87
10.2
4.92
(0.48)
2.66
(0.26)
-0.82
(-0.08)
3.48
(0.34)
2.66
(0.26)
This paper
Japan
Firm
1994-2001
2.1
1.20
(0.56)
0.33
(0.15)
-0.09
(-0.04)
0.42
(0.20)
Barnes, Haskell, and Maliranta (2001)
Finland
Firm
1987-92
5.4
-5.08
(-0.94)
6.37
(1.18)
2.86
(0.53)
France
Firm
1987-92
-7.7
-10.16
(1.32)
1.46
(-0.19)
Italy
Firm
1987-92
15.5
8.22
(0.53)
Netherlands
Firm
1987-92
2.7
UK
Firm
1987-92
-4.5
15.60
(0.56)
0.50
(0.02)
0.61
(0.29)
1.13
(0.53)
-0.52
(-0.24)
3.51
(0.65)
4.10
(0.76)
2.92
(0.54)
1.19
(0.22)
1.62
(-0.21)
-0.15
(0.02)
1.00
(-0.13)
0.92
(-0.12)
0.08
(-0.01)
2.17
(0.14)
3.57
(0.23)
-1.40
(-0.09)
5.12
(0.33)
5.43
(0.35)
-0.31
(-0.02)
4.16
(1.54)
-0.16
(-0.06)
2.46
(0.91)
-2.62
(-0.97)
-1.30
(-0.48)
0.16
(0.06)
-1.46
(-0.54)
-6.93
(1.54)
1.40
(-0.31)
-1.04
(0.23)
2.43
(-0.54)
1.04
(-0.23)
0.23
(-0.05)
0.77
(-0.17)
Notes: The entry and exit effects in this paper and in Ahn, Kwon, and Fukao (2004) include the switch-in and switch-out effects, respectively. Values in parentheses denote the share of each
effect in total TFP growth.
Figure 4.1 Diffusion Index of Business Conditions ("Favorable" minus "Unfavorable") in the Manufacturing
Sector: by Firm Size
60
40
0
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
20
Large
Small
Large minus Small
-20
-40
-60
Source: Bank of Japan 'Tankan (Short-term Economic Survey of Enterprises in Japan)'
Notes: The BOJ revised the Tankan from the March 2003 survey onwards. In the case of the December 2003 survey, both
the data based on the old format and the data based on the new format are available. Using these data we linked the
statistics before and after the revision.
Before March 2004, small firms are defined as firms with 50-299 employees and large firms are defined as firms with 1000 employees or
more. After March 2004, small firms are defined as firms capitalized at 20 million yen or more to less than 100 million yen and large firms
are defined as firms capitalized at 1 billion yen or more.
Figure 4.2 Labor Productivity in the Manufacturing Sector: by Firm Size, Logarithm of
Value Added (in Million Yen) per Worker
2.5
2
Small
1.5
Medium and
Large
1
Medium and
Large minus
Small
0.5
2001
2002
2003
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
0
Source: Ministry of Finance Statistics Monthly: Annual Financial Statements Statistics of Corporations ,
Notes: Small firms are defined as firms capitalized at 10 million yen or more to less than 100 million yen.
Medium and large firms are defined as firms capitalized at 100 million yen and more.
Table 4.1 TFP Level Gap between the 75 Percentile Firm and the 25 Percentile Firm; by Industry and by Year
Industry
1994
1995
1996
1997
1998
1999
2000
2001
Change of
dispersion:
1994-2001
Average market
share of each
industry: 19942001
Drugs and medicine
0.158
0.165
0.180
0.201
0.197
0.207
0.227
0.216
0.058
0.021
Petroleum and coal products
0.123
0.121
0.142
0.197
0.164
0.155
0.193
0.176
0.053
0.030
Electronic data processing machines and electronic equipment
0.164
0.182
0.180
0.178
0.181
0.189
0.193
0.208
0.044
0.060
Electronic parts and devices
0.149
0.150
0.150
0.153
0.151
0.168
0.167
0.190
0.042
0.055
Miscellaneous transportation equipment
0.104
0.103
0.107
0.117
0.193
0.151
0.108
0.142
0.038
0.014
Other manufacturing
0.119
0.127
0.128
0.128
0.131
0.139
0.153
0.155
0.036
0.018
Oils and paints
0.094
0.088
0.098
0.092
0.101
0.101
0.145
0.129
0.035
0.010
Non-ferrous metals and products
0.100
0.085
0.096
0.093
0.097
0.104
0.117
0.122
0.022
0.024
Miscellaneous electrical machinery and supplies
0.132
0.120
0.115
0.127
0.117
0.141
0.138
0.154
0.022
0.013
Motor vehicles
0.100
0.099
0.101
0.103
0.106
0.114
0.114
0.120
0.020
0.135
Communication equipment and related products
0.167
0.183
0.156
0.154
0.181
0.156
0.166
0.186
0.019
0.037
Household electric appliances
0.124
0.116
0.114
0.110
0.128
0.142
0.153
0.136
0.012
0.008
Rubber products
0.122
0.121
0.124
0.122
0.115
0.128
0.138
0.131
0.009
0.010
Plastic products
0.106
0.096
0.109
0.100
0.105
0.111
0.115
0.115
0.009
0.027
Office, service industry and household machines
0.130
0.147
0.135
0.119
0.125
0.133
0.131
0.137
0.007
0.018
Fabricated metal products
0.123
0.119
0.112
0.107
0.116
0.129
0.129
0.129
0.006
0.053
0.024
Textiles
0.186
0.189
0.201
0.196
0.199
0.199
0.193
0.188
0.002
Industrial chemicals and chemical fibers
0.115
0.100
0.106
0.108
0.111
0.109
0.104
0.114
-0.001
0.042
Wood and furniture
0.112
0.099
0.099
0.097
0.095
0.094
0.110
0.109
-0.003
0.014
Special industrial machinery
0.120
0.124
0.112
0.111
0.106
0.117
0.135
0.117
-0.003
0.024
Pulp and paper
0.102
0.090
0.089
0.087
0.097
0.097
0.105
0.098
-0.004
0.025
Food
0.132
0.122
0.118
0.111
0.116
0.113
0.117
0.128
-0.004
0.119
Printing and publishing
0.157
0.141
0.145
0.137
0.136
0.141
0.153
0.151
-0.005
0.039
Miscellaneous machinery and machine parts
0.127
0.113
0.111
0.119
0.119
0.138
0.125
0.119
-0.008
0.043
Industrial electric apparatus
0.139
0.136
0.140
0.124
0.139
0.146
0.155
0.130
-0.009
0.030
Other chemical products
0.126
0.123
0.124
0.138
0.132
0.138
0.129
0.117
-0.009
0.017
Precision instruments
0.160
0.137
0.152
0.133
0.149
0.148
0.143
0.149
-0.011
0.016
Iron and steel
0.118
0.089
0.088
0.083
0.117
0.112
0.097
0.106
-0.011
0.038
Ceramics
0.140
0.123
0.132
0.123
0.123
0.114
0.124
0.126
-0.014
0.026
Metal working machinery
0.185
0.130
0.112
0.101
0.139
0.139
0.149
0.128
-0.057
0.009
Weighted average of all the industries
0.130
0.125
0.125
0.125
0.130
0.134
0.137
0.141
0.011
1.000
Textiles
Metal working
machinery
Ceramics
Iron and steel
0.06
Food
0.08
Printing and
publishing
Miscellaneous
machinery and
Industrial
electric
Other chemical
products
Precision
instruments
0.06
Industrial
chemicals and
Wood and
furniture
Special
industry
Pulp and paper
0.15
Communication
equipment and
Household
electric
Rubber
products
Plastic
products
Office, service
industry and
Fabricated
metal products
Motor vehicles
Non-ferrous
metals and
Miscellaneous
electrical
Oils and paints
0.4
Drugs and
medicine
Petroleum and
coal products
Electronic data
processing
Electronic
parts and
Miscellaneous
transportation
Other
manufacturing
Figure 4.3 Change in TFP Gap by Industry and Industry Characteristics: 1994-200
0.06
0.04
0.02
Change in TFP gap
0
-0.02
-0.04
-0.06
0.1
R&D intensity
0.04
0.02
0
0.8
0.6
Amount of materials purchased from abroad/total amount of materials purchased
0.2
0.25
0
0.2
Stock of direct investment abroad/total assets
0.1
0.05
0.08
0
Percentage of firms owned by foreign firms
0.04
0.02
0
Table 4.2. Panel A. Comparison of Firms' Average Characteristics between the Top Firm Group and the Bottom Firm Group: All Manufacturing Firms
1994－2001
rms' Average Characteristics between the Top Firm Group and the Bottom Firm Gro Below the Above the
t-test
25 percentile 75 percentile
A
B
0.019
0.044
***
Amount of materials purchased from abroad/total amount of materials purchased
0.004
0.010
***
Stock of direct investment abroad/total assets
0.277
0.390
***
Number of non-production workers/number of all workers
0.006
0.013
***
R&D intensity (R&D investment/total sales)
35.383
36.901
***
Number of years passed since being established
4.906
5.484
***
ln(number of workers)
0.803
0.642
***
Total liability/total asset
0.004
0.019
***
Number of firms majority-owned by a single foreign firm/total number of firms
0.218
0.356
***
Firms majority-owned by a single domestic firm/total number of firms
0.034
-0.021
***
TFP growth rate
0.020
-0.015
***
Growth rate of real sales
-0.031
-0.001
***
Growth rate of employment
0.013
0.083
***
Return on asset (ROA)
Gap
B-A
0.025
0.006
0.112
0.007
1.517
0.577
-0.161
0.015
0.138
-0.055
-0.035
0.030
0.070
Below the
25
C
0.021
0.004
0.285
0.006
33.964
4.932
0.821
0.005
0.221
0.045
0.044
-0.020
0.016
1994－1997
Above the
t-test
75
D
0.043
***
0.009
***
0.403
***
0.013
***
35.973
***
5.471
***
0.654
***
0.018
***
0.323
***
-0.017
***
0.013
***
0.006
***
0.087
***
Notes: *P=.10, **P=.05, ***P=0.01. In the t -tests heteroscedasticity among different percentile groups were taken account of.
Gap
D-C
0.022
0.005
0.118
0.007
2.009
0.539
-0.167
0.014
0.102
-0.061
-0.031
0.027
0.071
1998－2001
Below the Above the
t-test
25 percentile 75 percentile
E
F
0.018
0.046
***
0.005
0.012
***
0.269
0.378
***
0.006
0.014
***
36.824
37.692
***
4.880
5.498
***
0.784
0.631
***
0.003
0.020
***
0.215
0.389
***
0.025
-0.025
***
0.001
-0.038
***
-0.040
-0.007
***
0.011
0.081
***
Gap
F-E
0.028
0.007
0.109
0.008
0.868
0.618
-0.154
0.017
0.173
-0.050
-0.038
0.033
0.070
Table 4.2. Panel B. Comparison of Firms' Average Characteristics between the Top Firm Group and the Bottom Firm Group: by Industry
Drugs and Medicine
1994－2001
Variable
1994－1997
Below the 25 Above the 75
t-test
percentile
percentile
Amount of materials purchased from abroad/total amount of materials purchased
Stock of direct investment abroad/total assets
Number of non-production workers/number of all workers
R&D intensity (R&D investment/total sales)
Number of years passed since being established
ln(number of workers)
Total liability/total asset
Number of firms majority-owned by a single foreign firm/total number of firms
Firms majority-owned by a single domestic firm/total number of firms
TFP growth rate
Growth rate of real sales
Growth rate of employment
Return on asset (ROA)
A
0.027
0.002
0.474
0.036
42.719
5.083
0.689
0.017
0.182
0.032
0.027
-0.008
0.027
B
0.177
0.013
0.664
0.071
49.184
6.716
0.436
0.149
0.229
0.003
0.033
0.008
0.127
***
***
***
***
***
***
***
***
*
***
***
Gap
B-A
0.151
0.011
0.190
0.035
6.465
1.634
-0.253
0.132
0.047
-0.029
0.006
0.016
0.100
Below the 25 Above the 75
t-test
Percentile
Percentile
C
0.034
0.002
0.487
0.042
41.423
5.171
0.684
0.023
0.178
0.019
-0.002
-0.012
0.029
D
0.216
0.010
0.662
0.067
43.357
6.579
0.474
0.146
0.169
0.010
0.033
0.008
0.129
***
***
***
***
***
***
***
***
*
***
1998－2001
Gap
D-C
0.181
0.008
0.176
0.026
1.934
1.408
-0.210
0.122
-0.009
-0.009
0.036
0.020
0.100
Below the 25 Above the 75
t-test
percentile
percentile
E
0.027
0.002
0.450
0.029
44.730
4.928
0.708
0.009
0.180
0.042
0.050
-0.005
0.023
F
0.168
0.013
0.681
0.075
51.526
6.841
0.411
0.161
0.275
-0.003
0.033
0.008
0.129
***
***
***
***
***
***
***
***
**
***
***
Gap
F-E
0.141
0.012
0.232
0.045
6.796
1.913
-0.298
0.152
0.095
-0.045
-0.017
0.014
0.106
Notes: *P=.10, **P=.05, ***P=0.01 2. In the t -tests heteroscedasticity among different percentile groups were taken account of.
Non-ferrous Metals and Products
1994－2001
Variable
Below the 25 Above the 75
t-test
percentile
percentile
Amount of materials purchased from abroad/total amount of materials purchased
Stock of direct investment abroad/total assets
Number of non-production workers/number of all workers
R&D intensity (R&D investment/total sales)
Number of years passed since being established
ln(number of workers)
Total liability/total asset
Number of firms majority-owned by a single foreign firm/total number of firms
Firms majority-owned by a single domestic firm/total number of firms
TFP growth rate
Growth rate of real sales
Growth rate of employment
Return on asset (ROA)
A
B
0.028
0.007
0.198
0.006
36.302
4.924
0.768
0.012
0.285
0.028
0.004
-0.034
0.012
0.061
0.012
0.278
0.010
37.576
5.374
0.661
0.018
0.475
-0.024
-0.019
0.002
0.080
***
***
***
**
***
***
***
***
**
***
***
1994－1997
Gap
Below the 25 Above the 75
t-test
Percentile
Percentile
B-A
C
D
0.033
0.005
0.080
0.003
1.274
0.450
-0.107
0.006
0.190
-0.052
-0.023
0.035
0.067
0.025
0.007
0.198
0.004
34.337
4.915
0.787
0.009
0.286
0.034
0.032
-0.019
0.021
0.067
0.011
0.273
0.009
36.783
5.490
0.677
0.015
0.440
-0.028
0.008
0.009
0.078
***
**
***
***
**
***
***
***
***
***
***
1998－2001
Gap
Below the 25 Above the 75
t-test
percentile
percentile
D-C
E
F
0.042
0.004
0.075
0.004
2.446
0.575
-0.110
0.006
0.154
-0.063
-0.024
0.028
0.057
0.030
0.007
0.198
0.009
38.794
4.930
0.745
0.016
0.280
0.024
-0.018
-0.045
0.003
0.057
0.014
0.283
0.010
37.925
5.282
0.635
0.022
0.514
-0.020
-0.040
-0.005
0.082
Gap
F-E
**
**
***
***
***
***
***
***
***
0.027
0.007
0.085
0.002
-0.869
0.352
-0.110
0.006
0.234
-0.044
-0.022
0.041
0.079
Communication Equipment and Related Products
1994－2001
Variable
Below the 25 Above the 75
t-test
percentile
percentile
1994－1997
Gap
Below the 25 Above the 75
t-test
Percentile
Percentile
1998－2001
Gap
Below the 25 Above the 75
t-test
percentile
percentile
Gap
A
B
B-A
C
D
D-C
E
F
Amount of materials purchased from abroad/total amount of materials purchased
Stock of direct investment abroad/total assets
Number of non-production workers/number of all workers
R&D intensity (R&D investment/total sales)
Number of years passed since being established
ln(number of workers)
Total liability/total asset
Number of firms majority-owned by a single foreign firm/total number of firms
Firms majority-owned by a single domestic firm/total number of firms
TFP growth rate
Growth rate of real sales
Growth rate of employment
0.027
0.007
0.221
0.008
29.724
4.996
0.820
0.000
0.340
0.092
0.090
-0.034
0.050
0.014
0.415
0.021
36.618
5.841
0.682
0.005
0.478
0.019
0.052
-0.004
***
***
***
***
***
***
***
*
***
***
**
***
0.023
0.006
0.194
0.013
6.894
0.845
-0.138
0.005
0.137
-0.074
-0.038
0.029
0.022
0.007
0.216
0.008
29.627
4.950
0.814
0.000
0.309
0.091
0.132
-0.012
0.035
0.013
0.431
0.024
35.917
5.861
0.656
0.003
0.436
0.018
0.070
-0.003
0.013
0.006
0.215
0.016
6.290
0.911
-0.158
0.003
0.127
-0.072
-0.062
0.009
0.016
0.008
0.214
0.007
32.075
4.835
0.782
0.000
0.295
0.094
0.056
-0.052
0.057
0.015
0.424
0.023
36.599
5.990
0.685
0.010
0.527
0.019
0.037
-0.005
***
*
***
***
***
***
***
*
***
***
***
0.041
0.007
0.209
0.015
4.524
1.155
-0.096
0.010
0.233
-0.075
-0.019
0.046
Return on asset (ROA)
0.019
0.077
***
0.059
0.015
0.105
0.090
0.015
0.065
***
0.050
**
***
***
***
***
***
***
***
**
***
F-E
Electronic Data Processing Machines and Electronic Equipment
1994－2001
Variable
Amount of materials purchased from abroad/total amount of materials purchased
Stock of direct investment abroad/total assets
Number of non-production workers/number of all workers
R&D intensity (R&D investment/total sales)
Number of years passed since being established
ln(number of workers)
Total liability/total asset
Number of firms majority-owned by a single foreign firm/total number of firms
Firms majority-owned by a single domestic firm/total number of firms
TFP growth rate
Growth rate of real sales
Growth rate of employment
Return on asset (ROA)
Below the 25 Above the 75
t-test
percentile
percentile
A
0.021
0.010
0.297
0.014
28.483
5.033
0.761
0.002
0.243
0.019
0.084
-0.030
0.012
B
0.068
0.016
0.485
0.026
28.846
6.032
0.710
0.037
0.574
-0.031
0.050
0.000
0.074
***
**
***
***
***
**
***
***
***
***
***
1994－1997
Gap
B-A
0.047
0.007
0.188
0.012
0.363
1.000
-0.050
0.034
0.331
-0.051
-0.035
0.030
0.063
Below the 25 Above the 75
t-test
Percentile
Percentile
C
0.020
0.012
0.303
0.012
27.995
5.088
0.764
0.005
0.244
0.042
0.126
-0.022
-0.003
D
0.066
0.019
0.459
0.028
28.044
6.185
0.715
0.024
0.541
-0.007
0.118
0.012
0.097
***
*
***
***
***
***
***
**
***
1998－2001
Gap
D-C
0.045
0.007
0.155
0.016
0.049
1.097
-0.049
0.020
0.298
-0.049
-0.008
0.035
0.100
Below the 25 Above the 75
t-test
percentile
percentile
E
0.021
0.009
0.287
0.016
29.412
4.973
0.737
0.000
0.216
0.002
0.052
-0.037
0.001
F
0.073
0.016
0.504
0.025
28.422
5.897
0.717
0.039
0.627
-0.050
-0.003
-0.010
0.072
***
***
**
***
***
***
***
*
***
Gap
F-E
0.052
0.006
0.217
0.008
-0.990
0.924
-0.020
0.039
0.412
-0.052
-0.055
0.027
0.071
Electronic Parts and Devices
1994－2001
Variable
Amount of materials purchased from abroad/total amount of materials purchased
Stock of direct investment abroad/total assets
Number of non-production workers/number of all workers
R&D intensity (R&D investment/total sales)
Number of years passed since being established
ln(number of workers)
Total liability/total asset
Number of firms majority-owned by a single foreign firm/total number of firms
Firms majority-owned by a single domestic firm/total number of firms
TFP growth rate
Growth rate of real sales
Growth rate of employment
Return on asset (ROA)
Below the 25 Above the 75
t-test
percentile
percentile
A
0.025
0.007
0.175
0.008
24.865
5.173
0.861
0.013
0.397
0.059
0.092
-0.031
0.012
B
0.076
0.016
0.288
0.014
28.387
5.908
0.694
0.025
0.594
-0.013
0.028
0.006
0.102
***
***
***
***
***
***
***
**
***
***
***
***
***
1994－1997
Gap
B-A
0.051
0.009
0.113
0.006
3.522
0.735
-0.167
0.012
0.197
-0.073
-0.063
0.037
0.090
Below the 25 Above the 75
t-test
Percentile
Percentile
C
0.025
0.005
0.184
0.008
23.870
5.197
0.886
0.016
0.395
0.078
0.158
-0.005
-0.010
D
0.068
0.016
0.290
0.014
27.216
5.933
0.737
0.026
0.581
0.001
0.108
0.025
0.131
***
***
***
***
***
***
***
***
***
***
***
***
1998－2001
Gap
D-C
0.043
0.011
0.106
0.006
3.346
0.736
-0.148
0.010
0.186
-0.077
-0.050
0.029
0.141
Below the 25 Above the 75
t-test
percentile
percentile
E
0.026
0.010
0.162
0.008
25.834
5.126
0.833
0.012
0.428
0.046
0.045
-0.050
-0.008
F
0.073
0.015
0.290
0.014
28.380
5.942
0.687
0.018
0.648
-0.024
-0.028
-0.007
0.108
***
***
***
***
***
***
***
***
***
***
***
***
Gap
F-E
0.046
0.005
0.128
0.006
2.546
0.817
-0.146
0.006
0.220
-0.069
-0.073
0.043
0.116
Motor Vehicles
1994－2001
Variable
Amount of materials purchased from abroad/total amount of materials purchased
Stock of direct investment abroad/total assets
Number of non-production workers/number of all workers
R&D intensity (R&D investment/total sales)
Number of years passed since being established
ln(number of workers)
Total liability/total asset
Number of firms majority-owned by a single foreign firm/total number of firms
Firms majority-owned by a single domestic firm/total number of firms
TFP growth rate
Growth rate of real sales
Growth rate of employment
Return on asset (ROA)
Below the 25 Above the 75
t-test
percentile
percentile
A
B
0.011
0.006
0.225
0.005
35.292
5.015
0.821
0.003
0.262
0.019
0.028
-0.022
0.019
0.015
0.023
0.262
0.012
39.009
6.011
0.645
0.008
0.319
-0.024
-0.004
0.000
0.082
*
***
***
***
***
***
***
**
***
***
***
***
***
1994－1997
Gap
Below the 25 Above the 75
t-test
Percentile
Percentile
B-A
C
D
0.004
0.017
0.036
0.007
3.717
0.996
-0.177
0.005
0.057
-0.043
-0.032
0.022
0.063
0.011
0.004
0.238
0.004
32.969
5.044
0.846
0.003
0.260
0.012
0.029
-0.013
0.017
0.013
0.019
0.266
0.011
37.703
5.933
0.665
0.010
0.294
-0.038
0.001
0.003
0.109
***
***
***
***
***
***
*
***
***
*
***
1998－2001
Gap
Below the 25 Above the 75
t-test
percentile
percentile
D-C
E
F
0.002
0.015
0.028
0.007
4.733
0.889
-0.182
0.006
0.034
-0.050
-0.028
0.016
0.092
0.009
0.007
0.211
0.006
36.982
4.973
0.803
0.002
0.252
0.025
0.027
-0.028
0.015
0.017
0.027
0.257
0.013
40.395
6.096
0.624
0.007
0.344
-0.013
-0.007
-0.002
0.073
Gap
F-E
**
***
***
***
***
***
***
***
***
***
***
***
0.008
0.020
0.046
0.008
3.413
1.122
-0.179
0.004
0.092
-0.038
-0.034
0.026
0.059
Table 4.3 Descriptive Statistics of the Variables Used for the Regression Analysis on the Determinants of TFP Growth
Sample size
Sample average
Standard
deviation
Minimum value Maximum value
109,882
-0.029
0.134
-4.511
1.450
84,923
0.004
0.091
-3.547
4.565
Domestic-affiliate dummy (majority-owned by a
single domestic firm)
109,882
0.278
0.448
0
1
Foreign-ownership dummy (majority-owned by a
single foreign firm)
109,882
0.008
0.089
0
1
Amount of materials purchased from abroad/total
amount of materials purchased
108,193
0.028
0.110
0
1
Stock of direct investment abroad/total assets
109,882
0.007
0.027
0
0.792
Number of non-production workers/number of all
workers
109,882
0.330
0.249
0
1
R&D investment/total sales
109,882
0.009
0.021
0
1.639
No. of years passed since being established
109,882
36.814
15.315
0
111
ln (no. of workers)
109,882
5.172
0.975
3.912
11.254
Total liability/total assets
109,882
0.726
0.302
0
11.653
TFP level
Growth rate of TFP level
Table 4.4 Determinants of TFP Growth: 1995-2001, All Manufacturing Firms
(1)
Dependent variable: ln(TFP)t-ln(TFP)t-1
ln(TFP)t-1
-0.2935
(-15.46)
(2)
***
(Domestic-affiliate dummy (majority-owned by a single domestic firm))t-1
(Foreign-ownership dummy (majority-owned by a single foreign firm))t-1
-0.3022
(-15.38)
0.0122
(12.44)
0.0229
(4.99)
(3)
***
***
(Stock of direct investment abroad/total assets)t-1
(R&D investment/total sales)t-1
(Dummy for firms which do not report R&D expenditure) t-1
(No. of years passed since being established)t-1
ln (no. of workers)t-1
0.0180
(10.18)
0.1545
(3.36)
-0.0039
(-4.32)
-0.0002
(-7.20)
0.0076
(15.41)
***
***
***
***
***
(Total liability/total assets)t-1
Intercept
Industry dummies
Year dummies
Number of observations
R-squared
Notes: The values in parentheses are heteroskedasticity-robust t-statistics .
*P=.10, **P=.05, ***P=0.01
-0.0305
(-8.63)
yes
yes
84923
0.1858
***
0.0194
(10.80)
0.1504
(3.28)
-0.0042
(-4.71)
0.0000
(-1.97)
0.0069
(15.08)
-0.0094
(-4.35)
-0.0270
(-8.86)
yes
yes
84923
0.1897
***
***
(Amount of materials purchased from abroad/total amount of materials
purchased)t-1
(Number of non-production workers/number of all workers)t-1
-0.2964
(-15.25)
(4)
***
***
***
**
***
***
***
0.0172
(5.29)
0.0191
(1.77)
0.0166
(9.77)
0.1401
(3.03)
-0.0035
(-3.98)
-0.0002
(-7.73)
0.0072
(15.34)
-0.0076
(-3.71)
-0.0219
(-7.48)
yes
yes
83494
0.1871
***
*
***
***
***
***
***
***
***
-0.3024
(-15.42)
0.0125
(13.29)
0.0186
(3.93)
0.0160
(4.76)
0.0319
(2.92)
0.0182
(10.41)
0.1440
(3.12)
-0.0039
(-4.40)
0.0000
(-1.91)
0.0065
(14.60)
-0.0095
(-4.47)
-0.0248
(-8.29)
yes
yes
83494
0.1905
***
***
***
***
***
***
***
***
*
***
***
***
Table 4.5 Estimated Values of Beta-convergence Coefficients: by Industry and by Period
(Ａ)
1995－1997
Industry
(Ｂ)
1998－2001
Coefficient of convergence Coefficient of convergence
All manufacturing
-0.386
***
-0.235
Food
-0.242
***
-0.171
Textiles
-0.278
***
-0.132
Wood and furniture
-0.278
***
-0.320
Pulp and paper
-0.381
***
-0.382
Printing and publishing
-0.566
***
-0.160
Industrial chemicals and chemical fiber
-0.228
***
-0.169
Oils and paints
-0.462
***
-0.116
Drugs and medicines
-0.066
-0.272
Other chemical products
-0.187
***
-0.166
Petroleum and coal products
-0.223
***
-0.070
Plastic products
-0.672
***
-0.257
Rubber products
-0.173
***
-0.566
Ceramics
-0.422
***
-0.324
Iron and steel
-0.211
***
-0.191
Non-ferrous metals and products
-0.639
***
-0.108
Fabricated metal products
-0.555
***
-0.315
Metal working machinery
-0.395
***
-0.322
Special industry machinery
-0.419
***
-0.379
Office, service industry and household machine
-0.240
***
-0.310
Miscellaneous machinery and machine part
-0.476
***
-0.334
Industrial electric apparatu
-0.369
***
-0.246
Household electric appliance
-0.340
***
-0.263
-0.492
***
-0.301
Communication equipment and related products
-0.169
***
-0.230
Electronic data processing machines and electronic equipmen
-0.447
***
-0.256
Electronic parts and devices
Miscellaneous electrical machinery and supplie
-0.340
***
-0.309
-0.328
***
-0.215
Motor vehicles
-0.353
***
-0.319
Miscellaneous transportation equipmen
Precision instruments
-0.306
***
-0.293
-0.253
***
-0.060
Other manufacturing
Notes: *P=.10, **P=.05, ***P=0.01
We estimated specification (4) of Table 4.4 for each industry in order to obtain the convergence coefficients.
***
***
***
***
***
***
***
**
***
***
***
***
***
***
**
***
***
***
***
***
***
***
***
***
***
***
***
***
***
Coefficient of
convergence
-0.377
***
-0.239
***
-0.262
***
-0.271
***
-0.380
***
-0.540
***
-0.227
***
-0.567
***
-0.079
*
-0.186
***
-0.168
***
-0.654
***
-0.200
**
-0.421
***
-0.207
***
-0.611
***
-0.543
***
-0.402
***
-0.412
***
-0.265
***
-0.471
***
-0.358
***
-0.347
***
-0.467
***
-0.180
***
-0.434
***
-0.345
***
-0.321
***
-0.337
***
-0.308
***
-0.235
***
(Ｃ)
1994-2001
Dummy for 1998-2001
period
-0.004
**
-0.017
***
-0.022
***
-0.019
***
-0.005
-0.023
***
-0.017
***
0.008
-0.012
-0.018
***
-0.045
***
-0.015
***
-0.027
**
-0.016
**
-0.017
***
0.031
***
-0.024
***
-0.062
***
-0.036
***
-0.021
**
0.011
**
-0.034
***
0.006
0.028
***
-0.001
-0.026
**
-0.055
***
0.015
***
0.019
***
-0.042
***
-0.025
**
Interaction term
0.137
0.066
0.116
-0.050
0.000
0.352
0.057
0.407
-0.169
0.018
0.069
0.375
-0.344
0.097
0.013
0.488
0.218
0.091
0.029
-0.032
0.132
0.112
0.090
0.141
-0.039
0.166
0.040
0.102
0.018
0.017
0.164
***
**
*
**
***
***
***
*
***
**
***
**
*
Table 4.6. Comparison of the Divergence and Convergence Effects: All Manufacturing Firms
Variable
Amount of materials purchased from abroad/total amount of materials purchased
Stock of direct investment abroad/total assets
Share of non-production workers in total workers
R&D intensity
Number of years passed since being established
ln(number of workers)
Total liability/total asset
Foreign-ownership dummy (majority-owned by a single foreign firm)
Domestic-affiliate dummy (majority-owned by a single domestic firm)
Summation of all the above divergence effects
Convergence effect (the average difference of TFP level*the estimated coefficient value
of the beta convergence)
Standard error of the random shock
Notes: We used coefficients of specification (4) in Table 4.4.
1994-2001
1994－1997
1998－2001
The gap
The gap
The gap
(The
between the
(The
between the
(The
between the
estimated top group estimated top group estimated top group
coefficient
and the
coefficient
and the
coefficient
and the
value)*gap
bottom
value)*gap
bottom
value)*gap
bottom
group
group
group
0.000
0.025
0.000
0.022
0.000
0.028
0.000
0.006
0.000
0.005
0.000
0.007
0.002
0.112
0.003
0.118
0.001
0.109
0.001
0.007
0.001
0.007
0.001
0.008
0.000
1.517
0.000
2.009
0.000
0.868
0.004
0.577
0.003
0.539
0.004
0.618
0.002
-0.161
0.003
-0.167
0.001
-0.154
0.000
0.015
0.000
0.014
0.000
0.017
0.002
0.138
0.001
0.102
0.002
0.173
0.011
0.013
0.010
-0.091
0.0797
0.300
-0.114
0.0828
0.295
-0.072
0.0758
0.305
Table 4.7 Transition Matrix of Firms' Rank: 1994-2001
TFPlevel
group in
1994
Lowest TFP group
2nd TFP level group
3rd TFP level group
4th TFP level group
5th TFP level group
6th TFP level group
7th TFP level group
8th TFP level group
9th TFP level group
Top TFP level group
Lowest
TFP group
34.0%
19.9%
12.7%
9.3%
5.8%
4.2%
3.7%
3.9%
3.3%
2.4%
TFP-level group in 2001
2nd TFP 3rd TFP
4th TFP
5th TFP
6th TFP
7th TFP
8th TFP
9th TFP Top TFP
level group level group level group level group level group level group level group level group level group
17.3%
10.3%
9.0%
7.2%
6.8%
4.1%
4.1%
3.3%
3.9%
18.8%
15.0%
12.8%
9.2%
7.3%
4.5%
5.1%
4.4%
3.0%
15.4%
14.9%
13.3%
11.4%
8.6%
9.5%
4.6%
5.5%
4.1%
12.1%
14.1%
14.9%
13.2%
11.6%
9.3%
6.3%
5.1%
4.0%
9.8%
12.7%
13.3%
14.1%
11.7%
12.0%
9.1%
7.0%
4.5%
9.6%
10.2%
11.2%
11.7%
13.9%
13.1%
10.8%
8.6%
6.8%
5.4%
9.0%
10.5%
10.3%
13.1%
14.3%
13.2%
11.8%
8.7%
4.4%
4.9%
5.7%
9.1%
11.6%
14.5%
17.1%
17.3%
11.5%
4.4%
5.1%
5.0%
7.9%
9.6%
10.4%
16.3%
18.6%
19.5%
3.2%
3.6%
4.2%
6.2%
6.1%
7.9%
13.4%
18.2%
34.9%
Table 4.8 Comparison of Firms' Employment Growth and Firm "Exits" between the Top Firm Group and the Bottom Firm Group: All
Manufacturing Firms
Below the 25 percentile in each industry
Above the 75 percentile in each industry
Number of workers
Number of workers
Growth rate of
Percentage of
Growth rate of
Percentage of
Fiscal year
All firms
the survivors' Firms "exited" workers in the
All firms
the survivors' Firms "exited" workers in the
total workers
firms "exited"
total workers
firms "exited"
a
b
c=b/a
a
b
c=b/a
1994
825336
86358
10.5%
2475820
120704
4.9%
(3396)
(556)
(3419)
(375)
1995
759593
-3.3%
86344
11.4%
2734449
-0.9%
141268
5.2%
(3559)
(536)
(3582)
(370)
1996
694034
-2.2%
86792
12.5%
2743078
-1.1%
151256
5.5%
(3519)
(523)
(3541)
(382)
1997
728367
-0.8%
84792
11.6%
2620956
-0.6%
143750
5.5%
(3491)
(533)
(3515)
(319)
1998
696177
-5.2%
68078
9.8%
2418887
-1.5%
182579
7.5%
(3479)
(503)
(3501)
(348)
1999
679635
-3.0%
131938
19.4%
2381912
-1.5%
168144
7.1%
(3423)
(768)
(3448)
(530)
2000
644740
-1.7%
69637
10.8%
2214509
0.4%
104466
4.7%
(3182)
(469)
(3206)
(303)
2001
651308
-6.2%
2262099
-2.9%
(3329)
(3354)
Average value
-3.2%
12.3%
-1.1%
5.8%
Notes: The values in parentheses denote the number of firms.